Why Does an Overhead Crane Girder Sag? Camber and Deflection

A loaded girder inevitably bends downward. Engineers call that elastic bend deflection. Under GB/T 3811, the acceptance limits are 1/700 of span for duty classes A1–A3, 1/800 for A4–A6, and 1/1000 for A7–A8. In real terms, a 20 m A5 crane may deflect at most 25 mm under rated load. Fabricators preset a reverse bend during manufacture, and that reverse bend is camber. It usually sits at 0.9/1000–1.4/1000 of span, and its job is to offset sag and keep the trolley from drifting. The girder is the most valuable single part of an overhead crane. Once camber and deflection make sense, the core of structural safety is in hand. And that core matters daily.

Camber and Deflection: Two Confused Terms

Deflection: the elastic bend under load

Deflection is the downward elastic deformation of the girder at rated load, measured against the unloaded state. The word elastic matters. In a healthy structure, the deformation returns to zero when the load comes off, and the girder recovers its shape. Bending on every lift and springing back on every unload is the designed condition, and the stiffness requirements of GB/T 14405 for general-purpose overhead traveling cranes build around this quantity.

Camber: the reverse arch made on purpose

Camber means the middle of the girder is deliberately built higher. The reference value sits around S/1000 of span. The normal range runs 0.9/1000–1.4/1000, and relevant standards control the tolerances. It serves two purposes. First, it offsets loaded sag, so the trolley rail tends toward level under load. Second, if a slight sag ever develops, camber keeps the trolley from rolling toward mid-span on its own.

The arch support in a new shoe makes a fair comparison. A flat arch makes walking tiring, and a girder that has gone flat, or reversed into a downward arch, lets the trolley drift while brake wheel and motor take pointless punishment. Camber relaxes a little over years of service, and that is normal. But the drift still shows up. The useful signals are the speed of relaxation and the amount left.

One more point of load logic is worth stating here: the girder is an elastic body rather than a rigid one. At the instant of loading, mid-span sags, the web dimples locally under each wheel point, and the two ends show a slight upward curl. Design rules constrain camber, deflection, and local web stiffness together, so the trolley rides a rail flat enough in every position. Once elastic deformation is accepted as inevitable, chasing a girder that never bends looks neither economical nor scientific. The real controls are how much it bends and whether it springs back. Both answers belong on paper.

The Numbers: How Much Should Your Girder Bend

Deflection limits follow duty class, and the higher the class, the stiffer the requirement.

For example, a 20 m class A5 bridge crane passes if mid-span deflection under rated load stays within 25 mm. At that scale the eye barely detects any movement, so judging structural trouble from a gut feeling that it looks bent is both unscientific and error-prone.

The static load test: a 1.25-times gate

For new builds, major overhauls, and periodic inspections, GB/T 5905 requires a static load test at 1.25 times rated load. The load then hangs suspended for the prescribed time, and after removal the structure must show no permanent deformation, no cracks, and no wrinkling or peeling paint. That test checks strength and residual deformation, a different question from elastic deflection at rated load. Both numbers, therefore, belong in the file.

The international comparison is worth knowing too. CMAA 70 sets a general girder stiffness limit around L/800 for bridge cranes, and AIST TR-13 tightens to L/1000 for metallurgical and process cranes, where FEM 1.001 gives systematic rules for load combinations. Export projects frequently write these figures into acceptance clauses.

An extra condition on the trolley rail

Mid-span deflection alone is not enough. GB/T 14405 also limits rail-joint height difference, lateral rail straightness, and gauge deviation between the two rails. Trolley wheel binding sometimes has nothing to do with the wheels, and it follows rail movement from a girder not stiff enough under load. Keeping deflection and rail survey data in one report removes half the later detours.

Four Real Causes of Excessive Sag

Undersized selection: running an A4 crane as A7

Enough rated tonnage does not guarantee enough structural life, because the same tonnage comes in different duty classes. Fatigue-life requirements for a low-frequency repair shop and a three-shift steel mill sit several grades apart. Put an A4 girder on an A7 station, and camber disappears at a visible pace as alternating stress accumulates.

Chronic overload and side pulling

Daily slight overload, dragging floor workpieces with the hook, and inclined side pulling all push the girder past design stress. Accident impacts can also cause local yielding in one event, and emergency stops, collisions, and a falling load caught at the last instant all count. The service record of many sagged girders contains one such big job that explains everything.

Fabrication defects and radiant heat

Excessive web waviness, incomplete weld penetration, and weld flaws all leave actual stiffness below the calculated value. In metallurgical shops, radiant heat changes material properties and stress distribution, so camber loss then accelerates markedly.

Natural relaxation past design age

Internal stresses in steel redistribute over time and weld residual stress releases, so camber decreases slowly on its own. On girders past ten years without periodic surveys, the real issue is usually that nobody ever measured. The change rarely happens overnight. But measurement always ends the argument. Then the file tells the truth. But only if measured yearly.

A machinery plant offered the textbook scene: a ten-year-old girder had lost its camber to outright sag. Under heavy load, a trolley parked at mid-span rolled toward the center by itself. The driver blamed the brake, and repeated adjustments changed nothing. But the brake was never the problem. The rail had become a slope, and no brake torque can hold back gravity on a slope. Gravity always wins.

These four causes often travel together. Undersized machines are the easiest targets for production-chasing, so overload and side pulling follow. Hot shops then speed up both fatigue and stress relaxation. A diagnosis should not fixate on one number. Load history, weld appearance, web waviness, and years of camber records need review together. Only then can anyone decide whether the girder is simply tired or has suffered an injury. A tired girder can face downgrading, while an impact-damaged one needs crack-risk management. The two treatments diverge completely.

How to Measure Without Dismantling the Girder

Level or total-station method

In the unloaded state, surveyors take elevation at mid-span and at both ends, then measure the same points again under rated load. After support settlement is deducted, the elevation difference is the deflection. Accuracy is high and the data stay traceable, so inspection bodies favor this method. Traceability matters in disputes.

Tight-wire and ruler method

A taut piano wire runs along the top cover plate as a reference, and rulers hung at mid-span and the quarter points give readings recorded unloaded and loaded. The equipment is simple and suits shop self-inspection, but the natural sag of the wire itself needs a correction.

Two interference terms to deduct

Rail slope and installation-reference deviation make up the first deduction, and then comes settlement of the legs or the building structure. On a gantry crane it is especially important to tell whether the girder bent or the leg sank. The more valuable habit is a trend file. Each measurement sits next to factory data and previous years. Slow change can be planned, while a sudden jump demands an immediate check for overload and weld trouble.

Three operational points come before the reading. Loads should be calibrated standard test weights or workpieces of known mass, so an overload deflection test does not become a second injury. Measurement points stay fixed at mid-span and quarter points every time, or the data lose comparability. Outdoor work avoids midday sun, because the temperature difference between the sunny and shaded faces produces a noticeable sideways bend in a box girder, so morning or overcast readings are the cleanest. With these details in the inspection procedure, a machine’s deformation curve stays documented for more than a decade.

What to Do When Deflection Is Out of Limit

Flame straightening and prestressed tensioning

For a sagged girder with sound material, flame straightening combined with prestressed tensioning can restore camber. The cost is roughly 20%–40% of a new girder, a typical range that moves with local conditions. Stay clear-eyed about one fact: straightening changes geometry, but the fatigue life already spent does not reset with it. Yet spent life never resets.

Reinforcement

Web stiffeners, added cover-plate section, and weld build-up can raise stiffness and capacity, but a qualified design institute must re-check any reinforcement scheme per GB/T 3811. Welding steel plates on by eye at site is not acceptable, because the added weight changes wheel pressure and building corbel loads too.

Downgrading or replacement

After a remaining-life assessment, downgraded duty is a common bridge measure: the girder moves from a high-frequency station to a low-frequency one. When permanent deformation, cracks, or severe fatigue damage are present, replacement is the responsible choice.

Does a straightening report with acceptable numbers equal another twenty years of life? Clearly not. Fatigue damage is irreversible. Geometry can come back, but the cyclic damage already inside the material cannot. Treating straightening as renewal and running the original duty at full load is the deepest possible misreading of the word acceptable.

A suggested decision sequence

When deflection looks abnormal, the order should be clear. First stop the suspect dangerous duty and re-measure to confirm. Then engage a qualified body for weld non-destructive testing and remaining-life assessment. Only then compare the economics of straightening, reinforcement, downgrading, and replacement. Straightening is cheapest but adds no life. Reinforcement raises stiffness and adds weight. Replacement costs the most but buys a complete fatigue cycle. Amortize the four options over their future service years and compare them against downtime losses. The conclusion is then usually far clearer than a gut call.

Why Yuzhong

Yuzhong was founded in 1978 in Changyuan, Henan, and now has 48 years of manufacturing history. It selects Q355B/C/D steel by service environment. Box girders use submerged-arc welding and integral CNC machining, and camber is measured on every finished unit. The data are filed and delivered with the machine, and at the selection stage structure follows the true duty class per GB/T 3811 and ISO 4301-1. A6/A7 high-frequency stations never take lower-class machines as a shortcut. The range covers 0.5–20 ton European-style single-girder cranes and 5–50 ton LHB explosion-proof double-girder cranes, plus the MG series gantry crane. FEM, ASME, and CE certification are in place, with exports to more than 120 countries, and shop inspection and third-party witness load acceptance by SGS and BV are available.