Services · Asset Integrity and Engineering Consulting
Shaft and Rotating Component Failure Analysis
A broken shaft carries the record of how it broke on its own fracture surface. The origin is readable, and it is usually not where the load was.
Request a QuoteOverview
TCR Engineering Services Pvt. Ltd. carries out shaft failure analysis at its NABL ISO/IEC 17025:2017 accredited laboratory in Navi Mumbai, India (certificate NABLT0726MH18640), on failed shafts, rotors, gears, couplings and other rotating components. The fracture surface is read first, then the microstructure, hardness, chemistry and design, against a group archive of 9,000+ documented failure investigations.
How Fatigue Failure of a Shaft Reads on the Fracture Surface
Most failed shafts fail by fatigue, and a fatigue fracture is a dated record of the event. The origin, the direction of crack growth, the number of initiation sites and the size of the final fast-fracture zone are all read from the surface before a single specimen is cut, which is why the fracture faces must not be cleaned, fitted back together or wire-brushed before they reach a laboratory.
- The origin: a single origin points at a stress raiser, a keyway, a fillet, a nick or a hard spot. Multiple origins around the circumference point at a nominal stress that was simply too high.
- Ratchet marks: the steps between adjacent origins. Their number counts the initiation sites, and a row of them says the stress concentration was severe.
- Beach marks: the macroscopic arrest lines. Their curvature says whether the loading was rotating-bending, reversed bending, torsion or a combination, and their spacing says whether the load varied.
- The final fast-fracture zone: its area against the section is a direct measure of how highly stressed the shaft was. A small final zone means a low nominal stress and a long crack life; a large one means the shaft was working near its limit.
- Striations under the SEM: the cycle-by-cycle record, which confirms fatigue and, where the spacing is measurable, bounds the growth rate.
A torsional fatigue crack runs at roughly 45 degrees to the axis and a rotating-bending crack runs normal to it. That single observation separates a drive-train problem from a bearing or alignment problem before any test is run.
The Damage Mechanisms We Separate
Naming the mechanism is the whole job. A shaft that failed by corrosion fatigue and a shaft that failed by plain fatigue call for different corrective action, and the difference is visible only in the evidence.
| Mechanism | What it looks like | What it usually means |
|---|---|---|
| Rotating-bending fatigue | Crack normal to the axis, beach marks concave towards the origin | Misalignment, an unbalanced rotor, or a fillet radius below drawing |
| Torsional fatigue | Crack at 45 degrees to the axis, often a star or spiral pattern | Cyclic torque, a drive-train resonance, or shock loading |
| Corrosion fatigue | Multiple origins, corrosion product in the crack, no clear single site | The environment, not the load, set the initiation life |
| Fretting fatigue | Initiation under a shrink fit, coupling hub or bearing seat, with red or black debris | Micro-movement at an interface that was assumed to be rigid |
| Stress corrosion cracking | Branched, largely brittle cracking with little section loss | A susceptible alloy, a tensile stress and a specific environment |
| Hydrogen embrittlement | Intergranular fracture, often delayed, frequently near a plated surface | Hydrogen from plating, pickling, welding or cathodic protection |
| Overload and single-cycle fracture | Shear lips and dimpled rupture, or cleavage in a brittle material | A load event outside design, or a material below specification |
| Creep and thermal fatigue | Cavitation on grain boundaries, or a network of surface cracks | Service above the temperature the material was chosen for |
Where the evidence supports more than one mechanism, the report says so and states what further evidence would separate them. A single named cause that the evidence does not carry is worth less than an honest short list.
What a Shaft Failure Investigation Covers
A shaft failure investigation establishes four things: where the crack started, the mechanism that grew it, the service and design conditions that contributed, and the corrective action that stops a repeat on the rest of the fleet. The fracture surface is documented and preserved before anything is sectioned, so the evidence survives any later review. Metallography, grain size, inclusion rating and hardness are reported to ASTM E3, ASTM E112, ASTM E45 and ASTM E10.
Chemical analysis, mechanical testing, metallography and hardness are performed under the NABL ISO/IEC 17025:2017 scope of certificate NABLT0726MH18640, which carries 1,483 accredited test items. The investigation that draws them together is an engineering opinion and is signed as one.
Where the Answer Has to Survive Scrutiny
A shaft failure report is rarely read only by the plant. It goes to an insurer, to the original equipment manufacturer, to a contractor whose workmanship is in question, and sometimes to a court. It is written to hold up in all four.
- Insurance and warranty claims: the mechanism and the timeline decide liability, and both come off the fracture surface.
- Referee and third-party appointments: where a supplier and a purchaser disagree, an accredited independent laboratory with no commercial interest in either answer is the route out.
- Repeat-failure investigations: the second failure of the same component is a design or a duty problem, and the two fracture surfaces compared side by side usually say which.
- Fleet decisions: one failed shaft raises the question of the others. In-situ metallographic replication lets the surviving units be read without being taken out of service.
Where a failure points at a wider population, the investigation hands over to remaining life assessment or to fitness-for-service to API 579-1/ASME FFS-1 by TCR Advanced Engineering, so the decision about the rest of the fleet rests on the same evidence.
Where Shaft Failures Start: Keyways, Fillets, Fits and Repairs
Shaft failure analysis at TCR Engineering, a NABL-accredited laboratory in Navi Mumbai, India, usually traces the origin to a geometric or surface feature rather than to plain bar. Pump, fan, compressor, gearbox, agitator, mill-drive and turbine-generator shafts concentrate stress at the same few features, and each feature points to a different correction.
| Origin | Why the crack starts there | What the correction usually targets |
|---|---|---|
| Keyway end | A sharp root radius and the abrupt change of section where the keyway runs out | The keyway profile and root radius on the drawing |
| Shoulder fillet | A fillet radius below drawing, or an undercut left by machining or grinding | The radius, the surface finish and any grinding burn |
| Shrink fit, coupling hub or bearing seat | Fretting under micro-movement at an interface designed to be rigid | Interference, surface treatment and assembly practice |
| Cross-hole or oil hole | A local stress raiser with an edge that was never deburred | Edge finish and the position of the hole |
| Weld repair, hard-facing or metal spray | A hard, brittle deposit zone or residual tensile stress put onto a rotating member | The repair procedure, or replacement in place of repair |
| Thread or circlip groove | A notch at the groove root carrying the full bending stress | Groove geometry and stress relief |
Where the lube oil was sampled before the failure, send the ferrography results with the shaft. Bearing and gear wear debris often records the misalignment or looseness that started the crack.
More in Asset Integrity
No published insight on this site bears directly on Shaft Failure Analysis yet. The 3 most recent Asset Integrity articles are below instead, and the full index carries all 93.
-
Acoustic emission testing of storage tanks and FRP pressure vessels: what is active now, and what it takes to find out
TCR Advanced Engineering runs acoustic emission testing on above-ground storage tanks and pressure vessels, including new FRP vessels under…
-
Remaining life of an ageing 220 kV transmission tower: the damage mechanisms that decide it and how they are found
An ageing 220 kV lattice tower is limited by zinc loss, stub corrosion, bolt and member damage and foundation distress. How each is found, judged and monitored.
-
Reformer tube fitness for service: Evolve by TCR opens a complimentary webinar on 25 September 2026
Watch the complete recording of the 25 September 2026 webinar, and see the two-day intensive programme in Vadodara on 29 and 30 October. Both address…
Frequently asked questions
How do you tell a fatigue failure from an overload failure on a shaft?
By the fracture surface. Fatigue leaves beach marks, ratchet marks and a distinct final fast-fracture zone, and under the SEM it leaves striations. An overload failure leaves shear lips and dimpled rupture across the whole section with no arrest marks. The two are separated before any specimen is cut.
Should the two halves of a broken shaft be fitted back together before sending them?
No. Fitting the halves together rubs the fracture faces against each other and destroys the origin, which is the single most informative area on the component. Pack the two halves separately, protect the fracture faces from contact and from moisture, and do not clean, wire-brush or degrease them.
What material and history do you need with the shaft?
The material specification and any mill certificate, the drawing with fillet radii and surface finish, running hours, the load and speed duty, the date and circumstances of the failure, and anything that changed before it. Where those are not available the investigation still runs; it simply carries wider bounds, and the report says so.
Is failure analysis covered by TCR's NABL accreditation?
The individual test methods are. Chemical analysis, mechanical testing, metallography and hardness are performed under NABL certificate NABLT0726MH18640, which carries 1,483 accredited test items. The investigation that draws those results into a cause is an engineering opinion, signed as one, and no accreditation body accredits an opinion.
How long does a shaft failure investigation take?
It depends on how much of the sequence the case needs and on whether long-duration tests are called for, and the timing is confirmed in the quotation. Where a plant is down, say so when the sample is booked, and the work is ordered to give the operational answer first and the full report after.
Can you examine the other shafts without taking them out of service?
Yes. In-situ metallographic replication to ASTM E1351 reads the microstructure of a component in place, with no cutting and no sample removal. Sister shafts can be replicated during a planned stop and read against a reference base of 100,000+ in-situ replicas held across the group.
Why do shafts keep breaking at the keyway or the shoulder?
Because those features concentrate stress, and a fatigue crack starts where the local stress is highest rather than where the nominal load is greatest. A keyway end with a sharp root, or a shoulder fillet below its drawing radius, cuts fatigue strength sharply, so the lasting correction is usually a drawing change rather than a stronger steel.
Can TCR investigate a shaft failure from a plant outside Maharashtra or outside India?
Yes. Failed components reach the Navi Mumbai laboratory from across India and from 15+ countries, and the group operates in India, Saudi Arabia, Qatar and Kuwait. Ship both halves of the shaft in separate dry packing with the drawing and the running history, and book the investigation before dispatch so the sample is expected.