Upset training gives pilots an in-flight safety net

 - November 27, 2007, 9:58 AM

Ask most professional pilots about either the USAir accident in Pittsburgh or the United Airlines crash in Colorado Springs, when the Boeing 737s flipped upside down before impact, and the discussion often focuses on whether it was wake turbulence, a roll cloud or a rudder hard-over that caused the crashes. Make a case for survivability in either of these two accidents and people will raise their eyebrows as if to say, “Are you kidding me?”
Few experts, except perhaps the folks at Boeing and Don Wylie from Aviation Safety Training (AST), believe the pilots could possibly have survived a rudder hard-over, if that was indeed the cause as the NTSB claims. USAir Flight 427’s old analog flight data recorder did not track rudder position. One of the questions that remain, however, is if these pilots could have recovered their airplanes. The simple answer is that they were never trained to recognize and recover from such an upset.
Aviators, especially those who climbed the professional ladder from the ranks of flight instructors, think they have a pretty firm grasp on aerodynamics. But after spending a few days in the company of Don Wylie, most experts will find themselves wondering how well they really paid attention during their early flight-training days.

Wylie, a former Air Force F-4 pilot in Vietnam, saw a parting of the waves, so to speak, after the USAir crash. “You only have to listen to the transcript to grasp the significance of what happened, no matter what turned the airplane over. The final words of the captain to the first officer are ‘Help me pull.’ The USAir Boeing was fully stalled when it crashed.”

Holding the wheel fully aft denied the crew any chance of a recovery. In the Colorado Springs crash, the 737 was in a steep spiral all the way to the ground. According to AST, loss of control in flight has contributed to more transport-category aircraft accidents than any other single cause in the U.S. and is a close second to controlled flight into terrain (CFIT) as a killer worldwide.

Wylie’s two-day course of ground and flight training–called the advanced maneuvering program (AMP)–makes it clear that an in-flight upset need not be a death sentence. AST employs 21 instructors, including Wylie, at its hangar at Houston’s Hooks Airport. I had the opportunity to attend a GIV simulator session at SimuFlite’s DFW training facility that Wylie taught as part of recurrent training for two General Electric pilots, Nick Esposito and Mark Ripa. The simulator session was followed by a trip to Hooks for a two-hour flight in a T-34 to see if I could transfer my classroom instruction into stick-and-rudder techniques that might save my life. I learned that while old flying habits can be tough to break, it can be done.

AST has trained more than 1,900 pilots in the advanced maneuvering program and counts among its clients such corporations as Abbott Labs, Merck and Morgan Stanley, in addition to the U.S. Army. Another customer is Bill McGoey, GE’s chief pilot of standards and training for a fleet of two BBJs, two GIVs, three Challenger 604s and two helicopters at the company’s Newburgh, N.Y. base.

“We’ve always thought upset training should be a regular part of our pilot recurrent,” McGoey said. He said this concept was reinforced when one of GE’s early hangar partners purchased a Decathlon aerobatic airplane and regularly trained crews in unusual-attitude flying. “Because there was a great deal of press about possible control issues on the [737, and by implication the] BBJ,” McGoey told AIN, “we wanted to make our passengers feel as comfortable as possible that we were properly trained for an upset. I found that selling this training was actually pretty easy. The pilots who have been through AST’s upset program thought it was some of the best training they’d ever experienced.”

Back to some aerodynamic discussions, the nucleus of Wylie’s training. The USAir 737 was no longer flying, and it would have been necessary to reduce the angle of attack (AOA) to make the wing fly once more. The problem was that to those pilots, the view out the window was not one they’d been trained to see and their “muscle memory,” as Wylie calls it, told them instinctively to pull to escape the ground. That turned out to be a fatal mistake.

Training Is Everything
Wylie fervently believes that with the right training, the USAir pilots would have had a chance to recover. But combine the untrainable startle factor of a real upset with most pilots’ passing, rote knowledge of unusual-attitude aerodynamics and the outcome might have been sealed.

According to Wylie, “We’ve been training a generation of pilots who don’t use the rudder and have become systems operators, simply relying on technology. Students also worry too much about complying with ATC instructions and maintaining altitude, even when the world is falling apart around them.” AST emphasizes this last point by the flight-recorder playback of a Boeing 747 that rolled over at 35,000 feet after the pilots failed to notice an increasing AOA following an engine failure. With the aircraft upside down, the pilots asked ATC– which was completely unaware of the problem–for a lower altitude rather than admit they were falling out of the sky.

Wylie says it takes some visionary thinking by pilots, and the blessing of management, to begin flying outside the standard 1G box most have been trained in. “Some students have an edge, though,” said Wylie, “such as taildragger pilots, people who are not afraid to kick the rudder pedals when necessary.” Additionally, he says helicopter, glider, ex-military and even remote control (RC) pilots do a pretty good job of recovering from severe upset maneuvers.

After watching thousands of pilots try to fly themselves out of the dark corners of their aircraft’s performance envelope, Wylie concluded, “Pilots fly like they’ve been trained.” Fighter pilots learn to fly their aircraft to its performance extremes to gain a tactical advantage over their enemy. This just happens to translate into a succession of abrupt, high-G maneuvers.

On the civilian side, recurrent training reinforces a reluctance to work the aircraft’s controls hard, even during an emergency, for fear of upsetting the folks in back. Wylie said it’s time pilots learned how to “spill a little coffee” to save their lives and those of their passengers. “I’ve watched properly trained pilots [in a simulator] recover a 777 at 250 feet agl when the upset occurred.”

But recovering from an upset means more than simply understanding aerodynamics, for just as the upset is occurring, at a time when the two people up front need to be completely on top of what is happening and how to escape, the human brain slips into “hyper-vigilance,” as anyone who has ever been part of an automobile accident will understand.

Events appear in slow motion, as a person’s vision begins to narrow and auditory exclusion commences. Adults scream for their moms and the brain reverts to any form of movement that it believes will save it from extinction. Unfortunately, that “muscle memory” can force pilots to pull when they should push, or to push the wrong rudder pedal at the wrong time.

Another issue that regularly arises in Wylie’s classes is how poorly some crews are prepared for an upset in normal flight. He includes a slide of a typical transport-aircraft cockpit, with the foot rests prominently displayed. While the ability to rest their feet on the instrument panel represents success to some pilots, Wylie worries that many crews might not understand how vulnerable sliding the seat back and putting their feet up leaves them if the airplane suddenly rolls upside down.

Comfort is one thing, but “the pilot should always have the ability to command full control of the yoke and rudder pedals whenever they are needed,” he said. “They should also be properly belted in.”

Wylie also worries that pilots are religiously trained to aim their aircraft around the sky following the flight director’s commands. He suggests all pilots review how to instantly unclutter their CRT display should a recovery become necessary, to avoid following a command that could lead to the ground. A company’s CRM procedures should dictate who turns off the flight director and yaw damper during an upset.
The Four Forces
Everyone with a pilot certificate has heard the discussion about the four forces that affect an aircraft in flight–lift, drag, gravity and thrust. But do we really have any practical knowledge in our heads other than those old drawings in the FAA Handbook with arrows emerging from the aircraft at odd angles?

Sure, there are 2Gs of force applied to an aircraft in a constant-altitude turn that must be balanced by an equivalent amount of lift, but what does it feel like? And what does it feel like when you enter a 2-G turn in one second rather than in the nice lazy fashion we’ve all been taught? And what about quickly rolling into a 70-degree bank that requires 3.3 Gs’ worth of lift?

The aerodynamic discussions in AST’s classes, especially those that focus on an aircraft’s resultant lift vector, make a pilot consider perspectives on flying they either forgot or must admit they never really understood in the first place.

Wylie calls an upset “an excursion that is both unexpected and unplanned for,” no matter what the attitude. And after training thousands of pilots, Wylie has seen some trends that show how little a pilot’s experience level means when the earth and the sky transpose. For instance, when the attitude indicator shows brown side up during an upset, most pilots will instinctively kick bottom rudder to try and get away from the ground, an act that almost guarantees the maneuver will terminate fatally.

If rudder induces yaw during a hard-over to the left, Wylie teaches rolling the ailerons right. According to the way Boeing builds its airplanes, he said, the 737 has enough aileron at any speed above stall to overcome a rudder hard-over, if the crew reacts properly. But pilots must know enough to “push” as they reach a high angle of attack.

Think how that integrates into all of our previous flight-training lessons. The airplane is rolling out of control because of a rudder hard-over, or a wake vortex, with insufficient aileron to recover. There is only one option left to regain control of the aircraft–reduce the angle of attack. That means lowering the nose to unload the wing.

Once pilots enter hyper-vigilance, however, most are unable to push on the yoke. But pushing on the yoke will immediately increase the power to weight ratio as the wing begins flying again, and that in turn means roll response to aileron input will also increase dramatically.

Finally, the ground-school sessions focus on AST’s guiding light for pilots who find themselves upside down and unable to think their way out of the problem: step on the sky! When the airplane rolls, push the rudder on the blue side of the ADI to find the quickest route back to level flight. Little thinking is needed.

If there is no blue showing, simply push the rudder in the direction of the ADI pointer to accomplish the same result. In these noncognitive events, as Wylie calls them, when it is life or death, pilots always seem to push the rudder away from the ground coming up, even if it is the rudder that brings the aircraft closer to the ground more quickly, which is why he calls it noncognitive.

Wylie emphasized again the importance of using the rudder smartly, although not aggressively. He prefers the term “managed application of rudder” during an upset rather than anything that sounds remotely like stomping on a pedal. Some of the big transport builders and operators around the world have made it clear that pilots should avoid using the rudder during an upset to prevent setting up the oscillation some claim tore the vertical stabilizer off American Flight 587 (an Airbus A300) shortly after takeoff from JFK. Wylie thinks that caution is misplaced, and he proved it during the flight-training portion of the course.

The GIV Simulator
The advantage of practicing extreme attitudes in a simulator is that you can’t hurt yourself. The disadvantage of trying these maneuvers in a simulator is that your brain instinctively knows you can’t hurt yourself either, which can render some portions of simulator training relatively ineffective. Unusual-attitude recovery during recurrent training in the simulator is normally accomplished at 15,000 feet, where many pilots feel like a hero if they successfully roll an airplane through an upset. But if that upset occurs a mile on final at 400 feet agl, everyone on board will probably become a statistic.

Wylie’s training combines aerodynamic theory with a trip through the simulator for some maneuvers to drive home the classroom points. McGoey said, however, “Most companies send their pilots out for two weeks of training each year, so adding another day or two for upset training can be a tough sell for some.” Although unusual-attitude training is required by the FAA during recurrent, pure upset training such as the AMP program is not.

GE was kind enough to let me look over Esposito’s and Ripa’s shoulders as they worked their way through Wylie’s upset training add-on to their own GIV recurrent. Esposito and Ripa are very experienced pilots, although relatively new to GE. Neither one has a military flying background.

Right after takeoff, ATC asked Esposito to turn the GIV left 40 degrees for traffic. As he entered the turn, ATC asked him to expedite, which prompted Esposito to increase his angle of bank. In a now steep turn close to the ground, the left engine quit, making for a pretty exciting ride as he attempted to maintain control of an aircraft that was now trying to roll over on its back.

Wylie had instructed Esposito not to use the rudder during recovery, but to fly with elevators and aileron only. At this low airspeed and with the ground coming up fast, the ailerons were not terribly effective as Esposito tried to roll the aircraft level. But Wylie told him to lower the nose–probably the last thing most pilots would try. As Esposito reduced the angle of attack and the ailerons became more effective, the Gulfstream began to roll level. Some expert piloting on Esposito’s part kept the ground at bay despite the EGPWS yelling at him every second or two. It was an impressive demonstration.

Wylie next had him set up for level flight at 20,000 feet and 280 knots. Esposito rolled into a few 90-degree banks but did not pull or push on the elevators as he turned. To demonstrate just how effective the rudder can be in controlling the lift vector and hence the aircraft, Esposito tried controlling the aircraft with the rudder alone. The airplane was slow to recover, but it was definitely under control.

He next tried with no controls at all, simply using differential thrust to point the airplane where we wanted it, a trick perfected by the United Airlines DC-10 crew in Sioux City in July 1989. The value of an alternative means to make the airplane fly was clear.

We also tried some upsets in which Esposito recovered by rolling through the upset (the recovery preferred by many pilots) rather than stepping on the sky. The results were dramatic. At the end of the roll-through, the GIV’s airspeed was beyond the redline and the G-meter said the wings had departed the airplane, despite how good it looked out the window. We also lost about 5,000 feet during the recovery. Wylie said the lack of a “wind in the wires” feel to the simulator is a liability in this kind of teaching.

Wylie had Esposito fly the aircraft near the top of its performance envelope, around 45,000 feet, to demonstrate more aircraft control issues. As the aircraft approached an intentional stall–something that is not too tough at this altitude, where all performance parameters begin to merge–Wylie asked him to try and maintain altitude, the way we’re all taught in training for stalls and recoveries.

When the stick shaker rattled the wheel, Esposito released some back pressure, which cost some altitude. As hard as he tried to coax the aircraft through the shaker and back to his original altitude, it was clear it was not going to happen. Wylie’s point was not even to try. Make the airplane fly by unloading it and simply realize you must trade altitude for control. Worry about ATC later.

The final trick Wylie showed everyone was from a five-mile final point with the GIV’s gear and full flaps down. Wylie simulated a wake-turbulence upset by rolling the aircraft over about 120 degrees and having Esposito recover.

At low speed and in a dirty configuration, Esposito quickly learned that when aileron and rudder are not effective enough, the only option is to push–or stop pulling so hard–on the yoke to get the airplane turned around. Wylie again mentioned his earlier 777 recovery example.

After some more pretty fancy flying on Esposito’s part, we landed safely after another upset even closer to the ground. Ripa then took the left seat and repeated the same exercises.

In all, each GE pilot probably spent an hour-and-a-half in the left seat watching demonstrations of the aerodynamics Wylie had taught in ground school before trying to turn their knowledge into action. Esposito reacted pretty positively to the training and added, “I think I need to use the rudder much, much more.” Next it would be my turn to integrate what I’d learned with my stick-and-rudder abilities in the Beech.

The T-34 Flight
Back in Houston, the midday July heat was already almost unbearable at Hooks Airport as Wylie and I preflighted Beechcraft T-34 N44KK. He showed me the basics of how to shed the aircraft’s clear bubble canopy should a problem arise, as well as how to pop out of the five-point harness and use the parachute should that become necessary–a checklist item that works wonders on the confidence of we straight-and-level guys. The T-34 uses a stick for control rather than a yoke and also has a G-meter to record the truth about past maneuvers.

Since AST likes to train T-34 pilots in pairs, we teamed up with student Charles Kerins, the owner-pilot of an Aztec based in the British Virgin Islands, and his instructor, Blake Thomas, in another of AST’s T-34s. Since both instructors are ex-fighter pilots, we taxied out in formation to the end of runway 17R. After a quick run-up, we took off together and turned right, northward toward Lake Conroe. Once we reached 3,500 feet and were clear of the Houston Class B, we tried some 60-degree-bank steep turns to feel the 2 Gs of back pressure necessary to maintain altitude. Wake turbulence is always a hot topic. With the smoke turned on from our formation’s lead ship, we flew through the wake of the other T-34. Even though these are small, relatively light aircraft, the roll motion generated as we flew through the smoke was eye-opening. Then the two T-34s went their own ways for separate training sessions.

Wylie and I quickly moved into accelerated stalls. But the way Wylie teaches them is totally different from what I recall being taught, or ever taught my own students. We were trained to fly them with power and ease the yoke back–albeit quite quickly–to feel the G forces and realize that an airplane stalls at a higher speed when the wing is loaded. When Wylie demonstrated the first accelerated stall to me, we pulled 2 Gs in about a quarter of a second and for the first time in my flying career I actually felt the G forces that make an airplane stall at a higher speed. My neck quickly succumbed to the compression of the stall, but we were out of the maneuver as quickly as we’d entered it.

I had Wylie try another so I could watch more closely, as well as feel it. This time it was clearly apparent not simply what 2 Gs felt like, but how instantly the aircraft began flying again when it was unloaded with the release of back pressure, just as I’d noticed during one maneuver in the GIV simulator. I tried a few of these myself and realized I had something new to teach my students.

Roll-offs were next as we flew through some simulated runaway nose-up trim maneuvers. Wylie had me push the stick forward to feel the force necessary to maintain level flight while he cranked in five degrees of nose-up trim. When I released the forward pressure, the Beech shot upward like a rocket.

Using ailerons and rudder only, I rolled the airplane into a 70-degree left bank and watched the climb stop instantly. The need for forward pressure on the stick immediately disappeared as Wylie’s explanation about the changing lift vector–now extending sideways–came back to me. The demonstration is designed to prove that the airplane–although erratic–can be controlled with a runaway trim before it rolls out of control. Forcing the stick forward to try to hold the airplane level could have exceeded its structural limits.

Another demonstration that would never have worked in the simulator is a wingover. It’s not that we couldn’t perform a wingover in the simulator, but we could never have recognized the nearly 4 Gs my ex-F-4 instructor pulled during the turn. Until a pilot has experienced this kind of G force, he’ll never completely understand how disorienting this seat-of-the-pants feeling can be. And we never actually rolled much past 120 degrees before recovering.

Then came the G demo, which will put the fear of God into any pilot who flies around with his feet planted firmly on the floor during cruise. Should a 2-G upset occur, something Wylie said is pretty easy, the pilot can’t even raise his legs off the floor or his arms to the yoke to pull–or push–as necessary against the force of gravity to regain control.

Now it was the day of reckoning, time to try stepping on the sky and see if it really made sense in flight. I thought I’d seen the wisdom of the technique during the GIV ride, but my guess was that from the way Wylie kept yelling to “push, push” on the rudder, I wasn’t quite fast enough. Wylie rolled the airplane into a right turn with about 160 degrees of bank–not quite upside down but close. I tried stepping on the sky as I pushed the stick to the left, but the airplane acted as if it were trying to decide what I’d asked it to do before it made a move. Wylie told me to “Push that left rudder next time.” I thought I had. Another 160 to the left and I still wasn’t fast enough to avoid watching the airspeed climb close to redline and almost force me to yank the stick back, pulling
a bunch of Gs.

Time to try another maneuver as I felt myself beginning to liquefy in the Texas heat. We moved to spatial disorientation to prove how easily even an experienced pilot can lose track of where he is when he begins to exit the warm, fuzzy 1-G world. I should have realized at this point that Wylie never gives students things they’ve seen before, although he’s equally diligent in trying to recognize a student’s limits. Scare him or make him sick and the learning is over–something he said he has witnessed with pilots of all experience levels.

Recoveries from steep turns with aileron alone prove the airplane can be coaxed back to level flight without using the rudder at all. But what is truly eye-opening is when recoveries from those 160- to 180-degree-back excursions are made using both aileron and rudder. The airplane almost jerks back to attention.

Wylie gave me another chance to step on the sky and this time, as we were rolled just about inverted, I fought my desire to roll through and pressed–hard–on the right rudder pedal as I pushed the stick right. The airplane did snap right back. “Now you’re talking,” Wylie chuckled. We tried a few more and the theory began to sink in.

The only item on the checklist we had no time for was spinning. We rejoined the other T-34 for landing. Once we returned to the hangar for a debrief, my wingman–or maybe I was his–Charles said, “The AMP worked for me. I froze the first time we tried a spin, though. If that had happened for real before the course, I guess I would have had no chance of recovery.”

No doubt most of the training AST offers will make many pilots uncomfortable, to say the least, from the ground-school statistics and cockpit-voice and flight-data recorder playbacks, to the simulator upsets that sometimes end with that deafening rumble and flashing of red lights characteristic of a simulated crash, to the loss of vision and feeling that your head is on its way to the bottom of the airplane as the blood races toward your feet in a real-life 3-G maneuver. But making people uncomfortable enough to be willing to spill the coffee in the back might just be the only way they’ll live long enough to talk about the value of this kind of training.

Admittedly, some of the maneuvers Wylie demonstrates are on the edge of statistical reality. But if the big one happens, this kind of training might be the only safety net a pilot owns. After a pilot completes the course, he’ll emerge with a healthy respect, as well as a newfound experience level, for how to survive an upset that will help offset the muscle memory that all pilots have used since their early training days. Aviation Safety Training’s AMP costs $1,000 per pilot as an add-on to a recurrent simulator session. The initial T-34 training costs $2,995, with recurrent training in the Beech set at $1,495.

Despite the fact that I nearly lost my lunch in the 98-degree Texas heat, I left Hooks Airport with a sense of awe that I’d just learned something I wish I’d known during the 34 years I’ve been flying. I think AST should try purloining and modifying the old American Express card slogan for its training: “Don’t leave the ground without it.”