paperburn1 wrote:A point I'd like to make it and I don't think a lot of non-aviation types understand is that how violent control reversal or flutter can be. It's not unheard of for a pilot to break his neck or otherwise severely injure themselves.
Quite an interesting point. Static and Dynamic aeroelasticity are unknown phenomena to most pilots and even some very talented instructors have limited knowledge about all the implications of bringing a plane to these limits.
This could as well been a case of "divergence" associated with a "control reversal situation". Considering also that they was flying in Trans-sonic regime ("between Mach 0.94 and Mach 1.02 that Michael Alsbury is seen on recovered cockpit video moving a lever to unlock the feathering system"), everything becomes more complex.
Lot of lessons will probably be learned from this accident.
hanelyp wrote:I am wondering if, with the feathering system unlocked, transonic flow may have overwhelmed actuators normally not engaged at that speed.
Yes, it could have. Testimony from the surviving pilot about craft behavior prior to structural collapse could clarify a lot.
The anonymous sources within Scaled Composites revealed that Siebold was able to unbuckle from his seat and deploy his chute at 17,000 feet (5,181 m). It is very likely that even Siebold fell unconscious from the initial stresses of the breakup and from decompression at 50,000 feet (15,240 m). He would have fallen into an unconscious state at that height and only have woken up once near 17,000 feet (5,181 m) where the atmosphere is denser and at which a human can survive, such as at mountain altitudes in the Andes and Himalayas. Whether he gave a thumbs up to a nearby chase plane is sensational but it would indicate that he was conscious and aware. With the parachute integrated into his test pilot suit, it was critical for Siebold to regain consciousness and unbuckle from his seat in order to give his parachute any chance of deploying. This is likely where the fate of the pilots differ.
Alsbury quite possibly was struck by debris or was injured by G forces and decompression more severely than Siebold. He either never regained consciousness or was somehow trapped in his seat and surrounding debris of the cabin. The circumstances for Siebold in his descent after the breakup were apparently fortuitous and gave him the chance to re-awaken and unbuckle. Comments in press reports from people around the incident or aware of the technology included that the pilots’ parachutes had automatic deployment mechanisms which activate at 10,000 feet (3048 m). In Alsbury’s or Siebold’s situation, without releasing themselves from their seats, the automatic deployment system would not have worked. If the chutes were to automatically deploy while the pilots were still strapped to their seats, the force from the deploying chute would have caused serious injury to the pilot.
I am not a nuclear physicist, but play one on the internet.
hanelyp wrote:I am wondering if, with the feathering system unlocked, transonic flow may have overwhelmed actuators normally not engaged at that speed.
Sure, or more likely made them unresponsive, but that's probably not the real issue. Check Yeager's biography regarding the X-1 controls. They became unresponsive as they approached Mach 1 in his early flights, but he overcame the problem by using the trim feature on the X-1, which was to adjust the angle of the entire horizontal stabilizer. It worked.
The result was switching to the "all flying tail" on the F-86, which solved a similar controls problem it suffered when going transonic. These are pretty standard today.
If the feathering system of the SS1 and SS2 craft is not an "all flying tail", nothing is. But it would be a flying tail with no control components.
But once unlocked, if the craft is not at an airspeed and attitude that lets it shift to its re-entry configuration, I see a pronounced opportunity for the tail to enter a non-linear oscillation, the end result of which I would expect to be the tail parting from the aircraft.
The data will tell. OSC probably already knows, but is letting the investigation play out so as not to ignore something even more important. Even unlocked, the tail should not have done what it did, and the second malfunction is the more important. The first can be fixed with training and checklist revision.
Another possible contributing factor, not yet mentioned, is Pilot-Induced Oscillation (PIO).
In the transonic regime, where this accident occurred, the susceptibility to PIO (and many of the other factors mentioned above) is higher than on either side of Mach 1.
Capturing all of the nonlinear system dynamics in a simulator (aeroelastic structural modes, uncertain transonic flow as shock waves form, hybrid motor thrust variations, liquid oxidizer slosh, etc.) is very difficult. Approximations and numerical table look-ups typically used in the simulator to model certain effects become less and less valid as dynamic ranges, number of active frequencies, cross-couplings, and various departures from linearity increase, reducing the predictive ability of the sim.
Perhaps an unfortunate combination of circumstances, a tragic synergy of dynamic "tolerance stacking" not seen in the simulator for the particular set of test cases that were run preflight.
EDIT: Hadn't read the interweb reports about insider info on the lever unlock at M1, until now. It seems more certain than I'd thought that it actually happened, so my comments above about PIO and "tolerance stacking" seem moot for this incident (still valid for some others).
It is true this happened on reentry, is it not? Just asking because there are no control surfaces up there. This is why the tail rotates--to act like a drag chute because there is not enough air to have control surfaces. Sounds to me like the drag chute was opened early and couldn't take the abuse.
"Courage is not just a virtue, but the form of every virtue at the testing point." C. S. Lewis
When they say the tail boom was "unlocked" early, at 1.0 gee rather than 1.4 gee, one supposes a little early, not an entire flight phase early. Could this really be what happened?
"Courage is not just a virtue, but the form of every virtue at the testing point." C. S. Lewis
"I would like to emphasize that what I'm about to say is a statement of fact, and not a statement of cause," Hart said, speaking at a late Nov. 2 press conference. There, he used two models of SpaceShipTwo to show how the vehicle "feathers" during reentry. Under normal operations, SpaceShipTwo is flown to a pre-determined altitude and released by its carrier aircraft, WhiteKnightTwo. For Friday's test, this happened at about 50,000 feet.
Shortly after SpaceShipTwo is released, it ignites its rocket engine and pitches upward, flying into a high arc designed to eventually reach the boundary of space. After the spaceship reaches apogee—the highest point of its flight—the vehicle's twin tails rotate upward, stabilizing and slowing SpaceShipTwo as it reenters the atmosphere.
It is not clear what the intended sequence of events was for Friday's test flight, but Hart said that the feathering system deployed prematurely. In order to feather the vehicle, two levers must be moved inside the cockpit. One unlocks the system, and the second deploys the stabilizers.
"About nine seconds after the engine ignited, the telemetry data showed us that the feather parameters changed from lock to unlock," Hart said. The telemetry data was confirmed by a camera inside the cockpit that showed co-pilot Asbury moving the handle. Normally, the feathers are not unlocked until the vehicle reaches Mach 1.4, Hart said. On Friday, Asbury unlocked the feathers around Mach 1.0.
However, neither pilot commanded the feathers to deploy. Shortly after the system was unlocked, the tail stabilizers began moving into their feather positions. Moments later, telemetry and video data stopped. The engine burn was normal up to the extension of the feathers, Hart said. This information runs contrary to speculation that the accident may have been caused by the engine, which was using a new fuel mixture during the flight. Scaled Composites, SpaceShipTwo's contractor, said the engine had been tested on the ground "many times," according to company president Kevin Mickey.
Hart said pieces of SpaceShipTwo, which are spread over a wide swatch of the Mojave Desert, are being moved into a hangar. "Among other things, [the investigators] found the fuel tanks, the oxidizer tanks, and the engine. All were intact and showed no signs of burn-through, and no signs of being breached," he said.
I am not a nuclear physicist, but play one on the internet.
From that, it sounds as if simply unlocking them early, while a breach of protocol, should not have changed the tail configuration. Perhaps a mechanical switch associated with the second lever was activated due to acceleration forces or perhaps some defect. A very unfortunate sequence of events that lead to the tragedy. But easier to fix than an engine/fuel problem. It is amazing that one pilot survived.
It is still hard to believe that any pilot familiar with this vehicle would intentionally arm the feathering mechanism prior to reaching a low dynamic pressure region, after motor burnout, near apogee.
If buffeting did not cause his hand to bump the unlock lever, my next suspect cause would be hypoxia. They were at 50,000 ft, after all.
I don't know how much difference it would have made in this case, but it makes sense to me that if instrument G forces approach safety limits the engine automatically throttles down.
The daylight is uncomfortably bright for eyes so long in the dark.
@JoeP
If the lever is unlocked near Mach 1 the normal aero forces alone are probably sufficient to overpower the feather actuators. The pilots would know this. Most primary flight control surface actuators (electric or electro-hydraulic) have modes such as floating-undamped, floating-damped, locked and active, all modes intrinsic to the actuator's local computer, but also selectable by the primary flight control computers. Since feathering is a secondary flight control function, it probably uses much simpler actuators with fewer modes, perhaps just on/off, with very little locking ability in the off state. I'm guessing that the lock function for SS2 feathering is provided solely by the lock/unlock lever in question, on the assumption that it would never be unlocked until dynamic pressure had dropped to where the actuators could hold until commanded to move.
@hanelyp
It's not clear if there is an autothrottle or not on SS2. If there is, then yes, it would make sense to throttle down when any of several parameters exceed safe bounds. But dynamic pressure killed this bird, and that is just 1/2 rho v**2, so chopping the throttle would probably be too little, too late.