Article by ALPA with NASA's help . Excerpts :
GPWS pull-up warnings require immediate crew action. Because most transport airplanes in CFIT accidents have hit within 200 feet of the top of the terrain, maximizing airplane performance at the beginning of the pull-up is crucial. The American Airlines B-757 that crashed on a ridge near Cali, Colombia, hit 100 feet below the ridge line--100 feet was the performance difference between survival and disaster.
A CFIT escape maneuver, a procedure usually performed in response to a GPWS warning, is designed to expediently remove an airplane from an impending collision with terrain. This maneuver is designed around the use of maximum or near-maximum aerodynamic performance. Typically, the airplane is in a descent. A CFIT escape maneuver can be started anywhere from a clean cruise descent (250 to 300 KIAS) to flying with full flaps and gear down at approach speed.
We looked at three cases: airplanes with conventional flight controls, FBW airplanes with "hard" protection features, and FBW airplanes with "soft" protection features.
Conventional flight controls
For an aircraft with conventional flight controls, the typical CFIT escape maneuver requires the pilot to apply maximum thrust and rotate at a smooth rate of 3 degrees per second to a pitch attitude of 15 to 20 degrees nose-up. This pitch attitude is maintained until the stickshaker activates or terrain clearance is ensured. The 3-degree-per-second rate was selected as a target pitch rate to avoid overstressing the airframe in the high-energy case and to avoid stalling in low-energy escape maneuvers.
An industry task force recommended this procedure, and it is the basis for most current CFIT escape maneuver procedures. Also, some operators had airplanes whose stickshaker may be activated by sensing AOA acceleration, and they believed a low-pitch rate was necessary to avoid activating the stickshaker. An operator selects for itself its CFIT escape maneuver procedure, and currently that may not be what the manufacturer recommends. Airlines often want to have a fleet-common procedure, regardless of aircraft type, for cross-training and standardization.
FBW flight controls
For the purpose of this investigation, we chose two airplanes that represent different FBW flight control design philosophies: the B-777 and the A330. We conducted test flights of the B-777 from Boeing Field in Seattle, Wash., and test flights of the A330 at the Airbus facilities in Toulouse, France. A company test pilot occupied the right seat in each case, and evaluation pilots took turns flying from the left seat. We operated both airplanes at a mid-CG with a takeoff weight that would permit the approach-configuration CFIT-avoidance maneuvers to be performed near the airplane's maximum (worst case) landing weight.
In preparing for the flight evaluations, we used the manufacturer's engineering flight simulator to work up a test card and to validate the profile. Simulator data allowed us to evaluate various maneuvers, determine the maneuvers with the best potential for optimal recovery, and practice the selected maneuvers.
Hard limits
Airbus, in the design of its FBW flight control system, incorporates "hard" limits, which prevent the pilot from exceeding a predetermined flight envelope. The FBW flight control system used in the Airbus design maintains an AOA margin to prevent the pilot from stalling the airplane. The aircraft cannot be commanded to exceed +2.5 gs or –1 g clean (regardless of gross weight). The pitch attitude is limited to the range between 30 degrees nose-up and 15 degrees nose-low. The bank angle is limited to 67 degrees. No approved or readily discernable method allows the pilot to override the flight envelope protections.
The A330-200's maximum takeoff weight is 507,100 pounds. For our flight, we weighed 402,600 pounds, just above the maximum landing weight of 396,800 pounds. Our fuel load was 134,000 pounds, with a maximum fuel capacity of 250,000 pounds. The A330-200's two P&W4168 engines are rated at 68,000 pounds of thrust each.
The procedure for Airbus's recommended CFIT escape maneuver in the A330 is for the pilot to pull full back on the stick, apply maximum thrust, and recover when terrain clearance is ensured. The speed brakes, if extended, will retract automatically. Control laws either stabilize the AOA at an optimum value or adjust pitch rate to obtain maximum allowed g.
Soft limits
Boeing, in the design of its FBW flight control system, incorporates "soft" limits, which warn the pilot when a limit is being approached, using such methods as increased stick forces and aural and visual warnings.
With soft limits, after the warning, the pilot is allowed to override and achieve maximum aerodynamic capability of the airplane; you could stall, overbank, overstress, or overspeed the airplane, if necessary or desired. The keyword for FBW soft limits is "overrideable."
The maximum takeoff weight of the B-777–300 is 660,000 pounds, and the maximum landing weight is 524,000 pounds. For our flight, the airplane weighed 501,000 pounds, of which 163,500 pounds was fuel. Two Rolls-Royce Trent 892 engines, each rated at 90,000 pounds of thrust, powered the airplane.
In a CFIT escape maneuver for the B-777, Boeing recommends that the pilot immediately select maximum thrust, rotate aggressively to 20 degrees of pitch, and retract the speed brakes (they do not automatically retract as do Airbus speed brakes), and recover when terrain clearance is ensured.
Flying the escape maneuvers
We flew simulated CFIT escape maneuvers near the maximum landing weight and at an approach speed of Vref +5 knots, with the gear down and flaps at maximum landing setting, at a descent rate of 1,500 fpm. We also flew simulated CFIT escape maneuvers at typical enroute descent speeds (250 and 300 KIAS as permissible), in the clean configuration, at a descent rate of 1,500 fpm. Splitting the task force's recommended 15–20-degree range for target pitch, we flew the maneuvers with a smooth pitch rate (target 3 degrees per second) to a pitch attitude of 17.5 degrees.
We then repeated these maneuvers with an aggressive pull-up (soft protections) or with full back stick (FBS) application (hard protections).
The benefit of the FBS maneuver for the A330-200 was quite apparent. Using full back stick resulted in an altitude loss of 40 feet, as opposed to 68 feet for the 3-degree-per-second recovery (see Chart 1). The evaluation team felt that time below entry altitude was as important a recovery parameter as total altitude lost. The time below the entry altitude was 3.4 seconds for the FBS versus 5.7 seconds for the aggressive pull-up. And finally, the FBS airplane was 172 feet above the 3-degree-per-second airplane, which was just getting back up to the entry altitude. Remember, most aircraft hit the ground within 200 feet from the top of terrain.
In the approach configuration, altitude loss during the FBS recovery was 35 feet, as opposed to 75 feet. The time below the entry altitude was 5.3 seconds versus 7.8 seconds (see Chart 2). And finally, the FBS airplane was 115 feet above the 3-degree-per-second airplane as the latter was just getting back up to the entry altitude.
As expected, data for the B-777 CFIT escape from an enroute descent showed that a rapid rotation method resulted in less altitude loss than did a slow rotation. Typical data showed about a one-third greater altitude loss with the 3-degree-per-second rotation. The test point showed 80 feet lost for 3 degrees per second (see Chart 4) versus 60 feet for the aggressive pull-up. The time below the entry altitude was also less, 3.1 seconds versus 4.7 seconds. The airplane with the rapid rotation rate was 180 feet above the slow-rotation-rate airplane as the latter passed through the entry altitude. Remember that in the Cali accident, 100 feet was the difference between clearing and not clearing the ridge.
The data for the B-777 configured for a power-on approach were interesting. The flight data showed a greater altitude loss for a rapid-rotation versus the slow-rotation maneuver--67 feet versus 50 feet shown (see Chart 3). Our simulator data predicted the opposite. While the data possibly have some scatter, the difference was attributed to the effects of the slow engine acceleration, which we first encountered while evaluating the airplane's stall characteristics.
The Trent engines demonstrated noticeably slow acceleration from low power settings, followed by a perceptible increase in engine noise and thrust surge as the engines "kicked in." After practicing, I could finesse the AOA rate of change during stall recovery to maintain optimum AOA during engine wind-up (I should mention, the test airplane had no dedicated AOA indicator).
In the low-speed CFIT escape maneuvers, however, the B-777's control laws apparently allowed overshoot of optimum AOA during this engine acceleration period, causing a "mush" effect that resulted in energy bleed and greater altitude loss. The 3-degree-per-second pitch-up was slow enough to allow thrust to catch up before the AOA increased to the same relative value.
We felt the Airbus pitch control laws prevented a similar AOA overshoot, even though the Pratt & Whitney engine acceleration times were similar. We did not see this effect in the high-speed escape maneuvers because enough energy remained so that AOA stayed low in both cases.
On the plus side, the B-777 rapid-rotation escape maneuver resulted in less exposure time below entry altitude than did the slow-rotation procedure (4.8 seconds versus 5.7 seconds), even though the altitude loss was greater. Passing through entry altitude, the rapid-rotation airplane was 50 feet above the airplane flown with the 3-degree-per-second rotation rate. On the line, different escape maneuver procedures might appear to be warranted for the high- and low-energy cases, with possibly a low-energy procedure addressing the Trent's slow spool-ups. As mentioned earlier, however, airlines tend to prefer "one size fits all" procedures. In addition, having an AOA indicator in the cockpit might help.
As in any flight evaluation, pilot comments are as important as the hard data. One of our most significant findings is that although we could achieve more consistent and repeatable performance with the "hard limit" design, this evaluation team philosophically preferred the flight envelope limiting features ("soft limits") of the B-777 design to the "hard limit" A330-200 design. This was a subjective judgment based on other handling quality evaluations that we performed and the premise that some situations might arise that the designers had not foreseen and for which the pilot might need to achieve full aerodynamic capability as opposed to being limited by software/control laws.
From our flight test work, we arrived at a number of conclusions and recommendations. First, the A330 full back stick CFIT escape maneuver gave better and more consistent performance than a 3-degree-per-second pull, without any increase in risk of exceeding flight envelope parameters. No additional or specific pilot training was necessary to perform the full back stick recovery technique because the FBW design provides excellent pitch rate and g control as well as excellent envelope protection for stall, overstress, or overspeed.
As a result, we recommend that Airbus FBW operators use the manufacturer's recommended full back stick CFIT recovery procedure. As I previously said, this may seem obvious; but until this report, none of the U.S. operators were following the Airbus-recommended procedure, and none felt that doing so was prudent. In addition, the ease of training and maneuver repeatability, in our opinion, outweigh the advantages of fleet-standardizing an airline's CFIT escape maneuvers for all airplane types.
In the case of the B-777, flight test results indicated that an aggressive pull-up as Boeing recommends yielded better CFIT avoidance performance than the 3-degree-per-second recovery procedure in all categories except total altitude lost in the case of the low-speed, low-power-setting situation. As a result, we recommend that B-777 operators follow the manufacturer's recommended CFIT escape maneuver procedure but consider modifying the procedure in the case of the low-speed, low-power-setting scenario.
As a final recommendation, based upon all of our previous conclusions, recommendations, and pilot comments, we feel that future FBW designs should consider protected flight-envelope limits with envelope-protection override.
I have flown all of the Airbus FBW airplanes and have been able to aggressively maneuver each airplane without worrying about overstressing the airframe. To pull the stick full back in an airplane that weighs half a million pounds and pull right to 2.5 gs is impressive (especially to the chief pilot if the airplane is loaded with passengers!). However, to be presented with a windscreen full of rocks and only a 2.5 g capability when more g is aerodynamically available is not comforting. This is when the capability to pull to the aerodynamic or structural limits of the airplane--as can be done in the B-777--is important. The B-777, however, does not have the carefree maneuvering capability of Airbus FBW airplanes. A combination of the best points from both these designs would be desirable.