Rotor Dynamics, Disks and Containment
The Jeffcott rotor and critical speeds, squeeze-film dampers and rotor modes, rotating-disk stresses and burst margin, and blade containment
United 232, 19 July 1989
A DC-10 flying from Denver to Chicago loses its tail-mounted engine: the stage-1 fan disk breaks apart. The DC-10's three hydraulic systems are each powered by one engine; debris penetrates the tail and severs the lines of the other two. With no flight controls, the crew — joined by a training check pilot travelling as a passenger, a throttle in each hand — steer with the two wing engines alone, and crash-land at Sioux City. Of the 296 people on board, 184 survive (United Airlines Flight 232).
The cause was a fatigue crack that had grown from an undetected metallurgical defect in the titanium-alloy disk. This lesson is about rotors that whirl, and the disks that must never break.
In 1869 W. J. M. Rankine analysed a spinning shaft and concluded that it could not run above its critical speed. In 1889 Gustaf de Laval ran a steam turbine above it anyway. The conflict lasted until Henry Jeffcott, commissioned by the Royal Society, published his analysis in 1919 (rotordynamics). His model — a disk on a flexible, massless shaft, its mass centre a small distance from its geometric centre — still opens every course on the subject.
The Jeffcott rotor: whirl radius r of the disk centre, and the angle φ by which it lags the mass centre, at spin speed Ω; ζ is the damping ratio.
Derivation
The Jeffcott rotor
Write the disk centre's position in the plane as a complex number . The mass centre is at , the shaft pulls the disk back with and damping resists with :
Try : the steady whirl radius and lag above follow, with . At only damping limits the orbit, . Far above it and : the disk spins about its own mass centre — the self-centring Rankine missed.
📐Through the critical speed
A 100 kg disk on a shaft of stiffness N/m — critical speed 4271 rpm — balanced to µm (Lab 1):
- Bare bearings, : the orbit peaks at 25 e = 500 µm, and each bearing feels 10 kN — enough to rub away tip clearances.
- A squeeze-film damper, : 51 µm and 1.1 kN.
- At twice the critical speed the orbit is back to 1.33 e, the high spot 176° behind the heavy spot.
💡Squeeze films, soft supports and rotor modes
A squeeze-film damper is a thin oil film around a bearing's outer race, held in a cage of springs: as the rotor whirls it squeezes the oil and dissipates energy. Engines combine it with deliberately soft supports, so that the rotor's first modes are rigid-body motions — bouncing and rocking on its bearings — at low speeds, passed quickly during start-up while the dampers soak up the energy. The rotor's own bending modes stay far above the running range. A mode's share of strain energy in the bearings tells which kind it is.
📐Stiff bearings or soft supports
A finite-element model of an HP rotor, its natural frequencies from
scipy.linalg.eigh (Lab 3), running range 9000–15,000 rpm:- Soft supports: modes at 3371 and 6131 rpm, with 99–100 % of their strain energy in the supports — rigid-body modes below idle, where the dampers work — and the first bending mode at 105,000 rpm, with 1 %.
- Stiff bearings: the first mode at 19,400 rpm, with its energy shared between rotor and bearings — above maximum speed, but with no damper able to reach it.
Derivation
Stresses in a rotating disk
A disk of thickness spinning at : radial equilibrium of a ring element and plane-stress elasticity, with radial displacement , give
The bore is free (); at the rim the blades pull outwards. Lab 2 shoots on the bore displacement to meet both. For a constant-thickness disk with a bore, the hoop stress at the bore is exactly
highest of anywhere in the disk and proportional to : the square of the rim speed.
📐Why turbine disks have fat hubs
A nickel-alloy HP turbine disk, 250 mm in radius, at 13,000 rpm with 70 blades (Lab 2):
- Constant thickness (75.8 kg): bore hoop stress 923 MPa, above the 800 MPa allowed here for fatigue life; burst at 1.44 times running speed.
- Tapered, four times thicker at the hub (166.2 kg): 593 MPa; burst at 1.70 times.
💡Blades are contained; disks are not
Certification requires an engine to contain the failure of its most critical compressor or fan blade, and of its most critical turbine blade, at maximum permissible speed: it must keep running for at least 15 seconds without catching fire or breaking its mounts (14 CFR 33.94). Engine cases are not designed to contain failed disks: the fragments carry too much energy. Disks are instead treated as safety-critical parts, designed, inspected and retired so that they never fail (turbine engine failure).
You want to be particular and make it a runway, huh?
⚠Southwest 1380
On 17 April 2018 a fan blade of a Boeing 737's CFM56-7B broke at a latent fatigue crack. The fan case was not penetrated — the blade was contained — but the impact broke the cowling latch, and cowling parts struck the wing and broke a cabin window. One passenger died in the decompression (Southwest Airlines Flight 1380). Containment covers the engine; the nacelle around it matters too.
💡Where you will use this
The same critical-speed maps, damper design and disk analysis apply to every turbine, pump and electric motor that spins fast. Lesson 9-3 follows the rotor's loads out through the frames and mounts.
📚Further reading
- John M. Vance, Fouad Y. Zeidan and Brian Murphy, Machinery Vibration and Rotordynamics (Wiley, 2010).
- Dara W. Childs, Turbomachinery Rotordynamics (Wiley, 1993).
- S. P. Timoshenko and J. N. Goodier, Theory of Elasticity (3rd ed., McGraw-Hill, 1970), on rotating disks.
↓But there is a problem
The rotors are held and the disks hold. But every load — thrust, weight, gyroscopic moments, a released blade — still has to travel out of the engine and into the wing.
On 4 October 1992, over Amsterdam, two engines left a Boeing 747.