Is that a reference to the one that destroyed a test-bed Vulcan (fortunately on the ground) when the shaft broke?
My source for that is Page 49 of "Modern Civil Aircraft: 2 Concorde" by Philip Birtles. He wrote that the cause was had been a difference in temperature in the shaft as the engine cooled after shut-down, leaving the top warmer than the bottom. When the engine was restarted and the power increased, the additional strain on the shaft that caused it to break. He wrote that the simple solution for Concorde's engines was a low rotation to even out the temperatures before restarting.
Ah no;- An engine comes up to temperature equilibrium very quickly typically 90 -120 seconds, maybe less on a smaller engine. That’s why you have a warm up time.
What happened was insufficient LP shaft structural damping in the torsion mode. What was “sufficient” wasn’t known at the time.
From “Sleeve Valve Notes”;- Rolls Royce Heritage & my dear friend G Ford (RIP) who witnessed this day by day.
The Oly 301 had an inter shaft bearing between the LP and HP shafts. While this worked it was troublesome leading short overhaul life’s and at least one major Vulcan incident(#). So on the 22R it was decided to delete the bearing, compensating for longer unsupported length by making the shaft much larger in diameter (increasing its section inertia). In addition it was decided to route cooling air through (maybe alongside) the shaft cavity.
Initially all went well on early testing, until out of the blue came a catastrophic failure on a ground test run in a test cell. The engine failed at full throttle and was so extensively damaged, determining it initial failure point proved extremely difficult. The failure was thought to have started in the new technology cast turbine blades so these switched back to tried an trusty forged blades. All was well when they started running again until a second, then a third failure in test cell and due to program pressure, the infamous Vulcan test failure that burnt out the fire engine.
Following an extensive investigation (•)it was discovered that a valve controlling air could flutter when near to closed;- a max power condition has minimal air bleed. This produced a cycling air pressure with the shaft causing to “ring like a bell”. The LP shaft should have dampened this out but was designed too close to its resonance harmonic so could expand and contract to the point it could touch the HP which it was running inside. Given the massive relative surface speeds a touch would be sufficient to sever the shaft, sheading all the turbine blades, releasing the turbine disc. In one early failure, the free disc ran around and around the test cell, like a buzz saw, cutting in two the overhead crane. Then the shattered engine couldn’t be removed but then again they couldn’t remove the broken crane either. With the Vulcan failed disc it bounced across the airfield for about a Km just missing the Bristol T188 being flight prepared.
The initial fix was to fully close the air valve and shrink a series of thin bands around the outside of the LP shaft. The final fix was the “Cake Stand”;- a series of disc mounted on a central shaft, heat shrink inside the LP shaft so that it considerable stiffened the shaft against bell resonance.
A very expensive test was undertaken at Pyestock whereby an engine was deliberately failed with a artificially induced pressure oscillation at the valve. Lots and lots of submarine netting limited the test cell damage, but still took about 18months to recommission the cell.
When TSR2 was flying they didn’t have the fully Cake Stand modified engines (external bands and air valve preloaded to shut only) so they installed a vibration pick up tuned to shaft resonance frequency wired to two lights in front of the pilot. They frequently illuminated but the shaft damped out successfully. I believe Roly got so fed up with them he ordered the lights be disconnected. Two fully Cake Stand modified 22Rs were either at Boscombe or about to be delivered when the project was cancelled.
The Cake Stand became a common feature on Bristol engines such as the Pegasus;- I’ve noticed recently it’s on a few non U.K. engines as well.
*a dynamics expert was brought in from NGTE Pyestock who was “an extraordinary clever chap”
# a Oly 301 was tested in a disused railway tunnel to see if they could get it to fail by cutting of the inter shaft bearing oil supply. It didn’t, so in addition to no oil they took the engine apart, sliced through part of the bearing support structure thus simulating a crack. Upon reassemble and test with no oil it properly went pop!
Edit
Just re-reading the article. They also shortened a aux pump drive shaft to decouple from the LP shaft and modified the reheat fuel flow.
To be fair there were a bunch of other problems on the engine;- bearing fires and fatigue cracking but pretty normal for early development and manageable.