Overhead Crane Bridge Girder Design Guide: Deflection Limits, Stress Calculations & CMAA Compliance
Overhead Crane Bridge Girder Design Guide: Deflection Limits, Stress Calculations & CMAA Compliance
Published by: WEIYUAN Engineering Team | Last Updated: July 2026 | Reading Time: 9 min
Article 1 of 3 in the WEIYUAN Overhead Crane Engineering series.
Introduction
The bridge girder looks like the simplest part of an overhead crane. It is a beam. It spans the bay, carries the trolley, and holds the load. Pick a section deep enough, weld the end trucks on, hang the hoist, and the crane works. That is how many people picture bridge girder design.
That is not how it actually works.
The bridge girder is the single most important structural component on the crane, and it fails in ways a static beam calculation never predicts. It carries dynamic loads that cycle hundreds of thousands of times over the crane’s life — not the one-time loads that govern most building beams. It must control deflection to a fraction of what a floor beam would allow, because a girder that sags lets the trolley roll toward mid-span on its own and turns every lift into a fight. And it must resist fatigue at the welds where stress concentrates, because that is where a working crane crane quietly seeds the cracks that end its life.
Here is how the cost of getting it wrong plays out. One buyer specs a girder to strength alone, meets the bending stress limit, and puts it into daily production. Within a few years the trolley is drifting on a visibly deflected girder, the wheels bind against the rail, and fatigue cracks appear at the girder-to-end-truck welds — a repair that means dropping the crane, cutting and re-welding primary structure, and losing the bay for a week. Another buyer lets deflection and fatigue drive the design from the start, pays a modest premium on steel, and runs the same duty for twenty-five years without touching the girder. Same span. Same load. The design method decided the outcome.
This guide covers the complete bridge girder design framework: the two girder types, the dynamic load inputs, the deflection limits that drive section selection, the section verification method, camber, the welded-versus-rolled decision, and the design errors that cost the most.
Part 1: Two Girder Types
Single Girder Cranes
In a single girder crane, one main girder spans the bay, and the trolley and hoist run on the bottom flange of that girder — the hoist hangs below, in an under-running arrangement. The end trucks sit at each end of the single girder and carry the whole assembly onto the runway.
Single girder cranes follow CMAA Specification No. 74 — the standard covering top running and under running single girder electric traveling cranes. They suit lighter capacities and shorter spans, typically up to around 20 tonnes and moderate spans, where the economy of one girder and a simpler build outweighs the reach and headroom advantages of a double girder.
The trade-off: because the hoist runs on the bottom flange, the girder sees bending combined with torsion and local flange bending from the eccentric trolley load — a more complex stress state that limits practical capacity and lift height.
Double Girder Cranes
In a double girder crane, two main girders run parallel across the bay, and the trolley runs on rails mounted on top of both girders — a top-running arrangement. The hoist sits between the girders, which raises the available hook height and lets the trolley carry heavier loads with better lateral stability.
Double girder cranes follow CMAA Specification No. 70 — the standard covering top running bridge and gantry cranes across the full range of industrial capacities, from light production cranes to the largest mill-duty cranes. They handle higher capacities, longer spans, higher duty classes, and give more room for walkways, festoon systems, and heavier trolleys.
The fundamental structural difference: the double girder carries the trolley load at the top flange through the rail, generating cleaner bending in the plane of each web, while the single girder carries the load eccentrically at the bottom flange, adding torsion. That difference is why double girder construction dominates as capacity, span, and duty class rise.
Part 2: Dynamic Load Inputs
Vertical Wheel Loads
The starting point for girder design is the trolley wheel load — the force each trolley wheel transmits to the girder as it carries the hoist and rated load across the span.
Maximum trolley wheel load = (Trolley dead weight + Rated load + hoist weight) × Dynamic impact factor ÷ Number of trolley wheels
The dynamic impact factor accounts for the shock of lifting, lowering, and traveling — the real force on the girder is higher than the static weight alone.
CMAA Specification No. 70 vertical impact (load) factors by duty class:
- Class A–B (infrequent to light use): approximately 1.10
- Class C–D (moderate to heavy production): approximately 1.15 to 1.20
- Class E–F (severe, continuous duty): approximately 1.20 to 1.25
Worst-Case Trolley Position for Bending
The bending moment in the girder is highest when the trolley sits near mid-span. At that position, the trolley wheel loads act at the point of maximum bending arm, producing the largest sag and the highest stress in the bottom flange.
For a girder span L with the trolley load P applied near mid-span, the maximum bending moment approaches PL/4 for a single concentrated load — the governing case for section strength. Both the moving trolley load and the girder’s own dead weight contribute to this moment, and both must be included in the design calculation.
Lateral (Side Thrust) Forces
CMAA Specification No. 70 requires the bridge girder to be designed for lateral forces acting horizontally at the top of the rail, perpendicular to the girder. These come from trolley acceleration and braking across the span, and from any skewing of the trolley on its rails.
Standard lateral force requirement: approximately 10 to 20% of the maximum trolley wheel loads (excluding the vertical impact factor), applied at the rail. The exact percentage depends on the duty class and the trolley’s acceleration characteristics.
This side thrust must be resisted by the horizontal stiffness of the girder — often provided by the box-section geometry itself, a horizontal flange plate, or a walkway truss — and transferred into the end trucks.
Longitudinal (Tractive) Forces
Bridge travel — the crane moving along the runway — generates longitudinal forces from acceleration, braking, and end-stop impact. These act along the runway direction and transfer through the end trucks into the runway and column bracing, rather than into the girder in bending. The design still accounts for them, typically at around 10% of the driven wheel loads.
Fatigue Load Cycling
Unlike most structural beams, which see their design load rarely, a bridge girder cycles on every lift. CMAA Specification No. 70 classifies cranes by load cycle class — from Class A at under 20,000 cycles to Class F at over 2,000,000 cycles over the design life.
This matters most at the welds. The girder-to-end-truck connections, the web-to-flange welds, and any welded attachments (rail clips, stiffeners, brackets) must be designed with fatigue in mind, not strength alone. AISC fatigue provisions (Appendix 3 of the AISC Specification) classify each weld detail by fatigue category, and the allowable stress range for each category falls as the number of cycles rises. A detail that is perfectly safe under static load can crack under two million cycles at a stress range the fatigue category does not permit — which is why high-duty girders use continuous welds and smooth, low-stress-concentration details at every connection.
Part 3: Deflection Limits — L/600 and Beyond
Why Girder Deflection Limits Are Strict
A floor beam might be designed to L/240 or L/360 under live load — limits chosen for occupant comfort and to avoid visible sag or cracked finishes.
A bridge girder is held to L/600 under CMAA Specification No. 70 for the vertical deflection from the trolley load — more than twice as strict as a typical floor beam. The reason is not appearance. It is how the crane works.
When a girder deflects at mid-span, it creates a shallow valley. The trolley, sitting on that slope, wants to roll downhill toward the low point on its own — so the operator fights the girder on every travel move, and holding the trolley at a position away from mid-span takes constant drive effort. On a top-running double girder, the same dip changes the wheel-to-rail contact as the trolley crosses it, raising rolling resistance and accelerating wheel and rail wear. Excessive deflection turns a smooth crane into a stiff, high-wear one.
For precision positioning applications — where the hook must land in the same spot regardless of where the trolley sits — L/800 to L/1000 is sometimes specified. That tighter limit all but eliminates any perceptible dip across the span.
The Deflection-to-Depth Relationship
Girder deflection under a given load is inversely proportional to the moment of inertia (I) of the cross-section, and I grows roughly with the cube of the girder depth for a given flange width.
The practical consequence is powerful: doubling the girder depth cuts deflection by a factor of around 8, for the same load and flange width. A modest increase in depth buys a large reduction in sag — which is why bridge girders are often noticeably deeper than a strength calculation alone would demand. On most spans, the deflection limit, not the bending stress limit, decides the section.
Span vs Depth — The Non-Linear Relationship
For a concentrated trolley load near mid-span, deflection is proportional to L³ / I; for the girder’s distributed dead weight, it is proportional to L⁴ / I. Either way, deflection grows far faster than the span itself.
So as span increases, the required depth has to grow faster than the span to hold I in proportion and keep deflection within L/600. This is exactly why long-span double girder cranes move to deep welded box girders — only a fabricated section gives the depth and the moment of inertia that a long span demands without becoming impossibly heavy. It is also why single girder cranes lose their economy at long spans and high loads: the section needed to control deflection outgrows what a rolled shape can practically provide.
Part 4: Section Selection Method
Every candidate girder section must pass three checks, in sequence. Skip any one and the section can look adequate on paper while failing in service.
Stage 1 — Bending Stress
Calculate the maximum bending moment from the worst-case trolley position (near mid-span) plus the girder dead weight, both with the vertical impact factor applied. Check the resulting bending stress against the allowable stress for the section.
For single girder (under-running) cranes, this stage must also account for lateral-torsional buckling and the added torsion from the eccentric bottom-flange load. For double girder box sections, the closed geometry resists torsion well, so bending in the plane of each web dominates.
Stage 2 — Deflection
Calculate the maximum vertical deflection under the trolley load (without the impact factor, per CMAA practice for the deflection check) and compare it against L/600 — or L/800 to L/1000 for precision applications.
In most practical designs, this is the stage that governs. The section that satisfies deflection is usually deeper and stiffer than strength alone requires, so Stages 1 and 3 become verification checks on the deflection-driven section rather than the drivers themselves.
Stage 3 — Local Effects
Global bending and deflection are not the whole story. The concentrated wheel load applies a local stress the global checks miss:
- Double girder (top running): web crippling and local flange bending directly beneath the rail, where the trolley wheel presses down. This is checked at the load point and controlled with web stiffeners where needed.
- Single girder (under running): combined bottom-flange bending and torsion from the trolley wheel running inside the flange, plus the flange-width check to keep the wheel tread off the flange tips.
A section that passes global bending and deflection can still develop fatigue cracks at the web-to-flange junction under repeated wheel loading if the local check is ignored.
Reference Sections by Span and Capacity
These are starting-point references for preliminary sizing only. Final selection requires the full load, deflection, and local-effect calculation for the specific crane.
- 5-tonne double girder, 10 m span: welded box girders roughly 550 to 700 mm deep
- 10-tonne double girder, 16 m span: welded box girders roughly 800 to 1000 mm deep
- 20-tonne double girder, 22 m span: welded box girders roughly 1100 to 1400 mm deep
- Single girder, up to ~10-tonne, ~12 m span: heavy wide-flange (W-shape) section with a reinforcing top channel or plate, or a light welded box
Part 5: Camber and Pre-Cambering
Why a Girder Is Built with an Upward Curve
A well-designed bridge girder is not fabricated straight. It is built with a slight upward curve — camber — along its length, so that under its own dead weight and a portion of the load it settles to level rather than sagging below it.
The reasoning is practical. A girder built dead straight will always sit slightly below the horizontal once its own weight acts on it, and lower still under load. That permanent sag creates the very trolley-drift and wheel-binding problems the deflection limit exists to prevent. Building in an opposing upward curve cancels the dead-load sag, so the finished crane presents a level track to the trolley.
How Much Camber
Typical practice is to camber the girder for the full dead-load deflection plus roughly half the live-load (trolley) deflection. A common reference figure is a camber of around L/1000 of the span, tuned to the calculated dead-load deflection of the specific girder.
The goal is a girder that reads as level — or very slightly crowned — under normal working conditions, never one that dips below horizontal. Too little camber leaves a residual sag; too much leaves a permanent hump that makes the trolley want to roll away from mid-span. The camber is matched to the calculated deflection, not guessed.
Building and Checking Camber
Camber is set during fabrication of the welded box girder — the web plates are cut to the cambered profile, and the welding sequence is controlled so weld shrinkage does not pull the built-in curve out. After fabrication, the camber is measured against the design profile before the girder ever ships, and it is checked again after installation with the crane unloaded. A girder that has lost its camber, or developed reverse camber (a downward sag) in service, is a warning sign of overload or fatigue and warrants a structural inspection.
Part 6: Welded vs Rolled Girder Comparison
The two ways to build a bridge girder — a fabricated welded box from steel plate, or a standard rolled wide-flange section — suit different capacities, spans, and budgets. Choosing the wrong one either overspends on a light-duty crane or under-builds a heavy one.
Welded Box Girder
A welded box girder is fabricated from four steel plates — two webs and two flanges — welded into a closed rectangular section, usually with internal diaphragms.
- Strengths: any depth and width you need, so it delivers the moment of inertia long spans and heavy loads demand; the closed box resists torsion and side thrust extremely well; camber is built in during fabrication; and it gives clean surfaces for rails, walkways, and festoon systems.
- Best for: double girder cranes, higher capacities, longer spans, and Class C–F duty — the great majority of industrial bridge cranes above light duty.
- The trade-off: more fabrication labor and weld quality control, so a higher cost per metre than a rolled section of similar depth.
Rolled Section Girder
A rolled girder uses a standard hot-rolled wide-flange (W-shape) section, sometimes reinforced with a channel or plate cap on the top flange for extra lateral stiffness.
- Strengths: lowest cost, fast to procure, minimal fabrication, and entirely adequate where loads and spans are modest.
- Best for: single girder cranes, light capacities, short spans, and Class A–C duty.
- The trade-off: limited to the depths and inertias available in the rolled range, so it runs out of stiffness at long spans and heavy loads, and it resists torsion far less well than a box.
2026 Indicative Cost Reference
Use these figures to budget the girder before detailed design. Costs are for the fabricated and coated girder, and vary with steel price, capacity, span, duty class, and coating specification.
| Girder type | Typical application | Indicative 2026 cost (USD) |
|---|---|---|
| Rolled W-section girder (with cap where needed) | Single girder, light duty, short span | $2,500 – $9,000 |
| Light welded box girder | Single or light double girder, moderate span | $6,000 – $18,000 |
| Standard welded box girder | Double girder, 5–20 t, medium span | $12,000 – $40,000 |
| Heavy welded box girder | Double girder, 20 t+, long span, Class E–F | $35,000 – $120,000+ |
The relationship holds across the range: a rolled section wins on price only where loads and spans stay modest, while a welded box earns its premium the moment deflection, torsion, or fatigue starts to govern — which is most of the time above light duty.
Part 7: Common Design Errors
Error 1: Designing for Strength, Not Deflection
The girder is sized to the bending stress limit and passes — but deflection is never checked, or is checked to a floor-beam limit like L/360. The girder is strong enough yet too flexible, so the trolley drifts toward mid-span, the wheels bind, and wear accelerates from the first shift. On most spans, deflection should have driven the section, not strength.
Error 2: Global Checks Pass, Local Wheel Stress Ignored
The section meets global bending and the L/600 deflection limit with the trolley load treated as a simple point load. But the concentrated wheel load creates local flange bending and web crippling directly beneath the rail — a separate check entirely. The girder passes globally and still develops fatigue cracks at the web-to-flange junction under repeated wheel loading, because the local effect was never verified and no stiffeners were added.
Error 3: Fatigue Category Overlooked at the Welds
The girder is designed for a single peak load and detailed with whatever welds are convenient. On a Class E or F crane cycling toward two million times, those weld details sit in a fatigue category whose allowable stress range is far below the static allowable. Cracks initiate at the girder-to-end-truck welds and attachment welds within a few years — a strength-adequate girder failing purely on fatigue because the AISC fatigue provisions were never applied.
Error 4: Camber Omitted or Wrong
The girder is fabricated straight, or cambered by a rule of thumb unrelated to its actual dead-load deflection. Built straight, it sags below level under its own weight and worse under load. Over-cambered, it holds a permanent hump. Either way, the finished crane never presents the level track the deflection limit was meant to guarantee, and the trolley-drift problem returns despite a section that passed every stress check.
Frequently Asked Questions
Q: Should I choose a single girder or a double girder overhead crane?
A: The decision comes down to capacity, span, duty class, and required hook height, not simply price. A single girder crane, with the hoist running under the bottom flange per CMAA Specification No. 74, is the economical choice for lighter capacities (broadly up to around 20 tonnes), shorter spans, and lighter duty (Class A–C) — it uses one girder, weighs less, and costs less to buy and install. A double girder crane, with the trolley running on rails on top of two girders per CMAA Specification No. 70, is the right choice as capacity, span, and duty class rise: it handles heavier loads and longer spans, gives noticeably more hook height because the hoist sits between the girders rather than below them, resists side thrust and torsion far better through its top-running geometry, and leaves room for walkways, heavier trolleys, and festoon systems. The practical rule is that a single girder wins on cost where the duty is genuinely modest, while a double girder earns its premium the moment you need real capacity, long span, high duty cycles, or maximum lift height — and above light duty, the double girder is usually the sounder long-term investment.
Q: Why is the bridge girder deflection limit (L/600) so much stricter than a normal building beam?
A: Because a bridge girder is not just holding a load — it is a running track for the trolley, and a floor-beam deflection limit would ruin how the crane works. A normal floor beam is designed to L/240 or L/360, limits chosen for comfort and to avoid visible sag. A bridge girder is held to L/600 under CMAA Specification No. 70 for one reason: when the girder deflects at mid-span it creates a shallow valley, and the trolley sitting on that slope wants to roll downhill toward the low point on its own. The operator then fights the girder on every travel move, and on a top-running double girder the same dip changes the wheel-to-rail contact as the trolley crosses it, raising rolling resistance and accelerating wheel and rail wear. So the tighter limit exists to keep the track effectively level, the trolley controllable, and wear low — which is also why precision applications go tighter still, to L/800 or L/1000, where any perceptible dip would spoil positioning accuracy. Practically, this strict limit is why deflection, rather than bending stress, usually governs the girder section on all but the shortest spans.
Q: Can I reduce the girder cost by using a rolled steel section instead of a fabricated welded box?
A: Sometimes — but only where the duty genuinely justifies it, and forcing a rolled section onto a heavy or long-span crane costs far more than it saves. A rolled wide-flange section, sometimes reinforced with a top channel or plate cap, is the cheapest girder to buy and the fastest to procure, and it is entirely adequate for single girder cranes at light capacities, short spans, and Class A–C duty. Its limitation is stiffness and torsional resistance: a rolled shape only comes in the depths and moments of inertia the mill produces, so at longer spans and heavier loads it simply cannot deliver the moment of inertia needed to hold deflection within L/600, and its open section resists side thrust far less well than a closed box. A welded box girder, fabricated from plate to any depth you need, gives the inertia long spans demand, resists torsion through its closed geometry, and lets camber be built in during fabrication — which is exactly why it dominates double girder and higher-duty cranes despite its higher cost per metre. The right approach is to let the load, span, and duty class choose the girder type: use a rolled section only where the calculation shows it comfortably meets strength, deflection, local wheel stress, and fatigue, and step up to a welded box the moment any of those starts to govern.