Accident at Roswell, New Mexico, 2 Apr 2011 (G650)

Accident narrative

On April 2, 2011, about 0934 mountain daylight time, an experimental Gulfstream Aerospace Corporation GVI (G650), registration N652GD, crashed during takeoff from runway 21 at Roswell International Air Center in Roswell, New Mexico. The two pilots and two flight test engineers aboard were fatally injured, and the airplane was substantially damaged by impact forces and a postcrash fire.

### The flight

The flight was operated by Gulfstream under 14 Code of Federal Regulations Part 91 as part of its G650 flight test program. Visual meteorological conditions prevailed. The test team was performing field performance flight testing to gather data for type certification and to develop takeoff and landing speed schedules.

The airplane’s angle-of-attack (AOA) limiter function was disabled because its software was not completed. Gulfstream intended for the stick shaker to provide a tactile warning of an impending stall, and for the pitch limit indicator (PLI) on the primary flight display to provide a visual indication.

The pilot-in-command (PIC) had 11,237 hours of total flight time, including 263 hours in the G650. The second-in-command (SIC) had 3,940 hours of total flight time, with 140 hours in the G650. Two flight test engineers (FTE1 and FTE2) sat at workstations in the main cabin. Five other test team members—a third flight test engineer (FTE3), three airplane performance engineers, and a telemetry engineer—monitored the flight from a telemetry trailer near the runway.

The team was conducting one-engine-inoperative (OEI) continued takeoff tests with the flaps set to 10 degrees. The G650 program had a takeoff performance guarantee of 6,000 feet at standard sea level conditions. Achieving this guarantee required the airplane to meet target takeoff safety speeds (V2) at a height of 35 feet. During previous tests, the airplane had consistently exceeded the target V2 speeds. To resolve these V2 overshoots, the test team had been experimenting with different takeoff rotation techniques.

During preflight briefings, FTE1 indicated that the target pitch attitude for the tests would be reduced from 10 degrees to 9 degrees. He also briefed that they should discontinue a test if pitch reached 11 degrees during the initial takeoff. The target V2 speed for the accident test run was 135 knots.

### The sequence of events

The test team completed 10 test runs earlier that morning. The on-board audio and video recorded the crew exploring rotation techniques to prevent V2 overshoots. The PIC noted that an abrupt "jerk" on the control column did not work and expressed concern about whether the Federal Aviation Administration would accept such a technique. The crew progressively modified their inputs to a "nice smooth ramp" of 50 to 55 pounds of column pull force, but they found that to keep the speeds down, they had to increase pitch above the 9-degree target progressively earlier after rotation.

At 0931, the Roswell air traffic control tower reported wind from 170 degrees at 9 knots. FTE3 in the telemetry trailer told the crew the wind was from 156 degrees at 5 knots, and the PIC replied, "we are okay with where we're at."

The accident takeoff, designated test run 7A2, began at 0933:00 when the PIC advanced the thrust levers. Between 0933:36 and 0933:37, at an airspeed of about 105 knots, the SIC moved the right thrust lever to idle to simulate an engine failure. At 0933:46, with airspeed at 127 knots, the SIC called "standby, rotate," and the PIC pulled the control column.

At 0933:50, the pitch attitude and AOA reached about 10 degrees, and the PLI appeared. The right main landing gear lifted off the runway. At 0933:50.4, pitch and AOA exceeded 11 degrees. The PIC stated an unintelligible remark as the bank angle increased to the right, and he made a left control wheel input.

At 0933:52.2, the stick shaker activated for 0.6 seconds. The SIC and PIC repeatedly said "whoa." The PIC pushed the control column forward abruptly and applied full left control wheel and rudder. At 0933:53.6, an electronic voice announced "bank angle" as the right bank reached 16.2 degrees. The PIC called for power, and the right thrust lever was advanced.

The right wingtip contacted the runway at a roll angle of 13.4 degrees. The stick shaker activated again and continued for the remainder of the flight. The PIC pulled back on the control column with more than 60 pounds of force, and the pitch and AOA increased. The bank angle reached 32 degrees right wing down. The airplane departed the right side of the runway, skidded across a taxiway intersection, and impacted a runway sign, a concrete structure, and an airport weather station. The impact caused extensive structural damage and a postcrash fire that completely consumed the fuselage and cabin interior.

### What the investigation found

The National Transportation Safety Board's aircraft performance study found that the airplane stalled while lifting off the ground. The stall occurred on the right outboard wing at an AOA of 11.2 degrees, producing a rolling moment that the flight crew could not control.

The investigation found that Gulfstream had overestimated the airplane's stall AOA while in ground effect. Ground effect increases lift and reduces drag, but it also reduces the stall AOA compared to free air. Gulfstream predicted the in-ground-effect stall AOA would be 13.1 degrees and set the stick shaker activation threshold at 12.3 degrees. Because the actual stall occurred at 11.2 degrees, the flight crew received no tactile or visual warning before the stall. The Board concluded that given the airplane's low altitude, the time-critical nature of the situation, and the ambiguous stall cues presented in the cockpit, the flight crew's response to the stall event was understandable.

The Board noted that this was the third time a right outboard wing stall occurred during G650 flight testing. Two previous uncommanded roll events had occurred during liftoff. On November 16, 2010 (Flight 88), an uncommanded roll was attributed by the test team to an over-rotation by the flying pilot. On March 14, 2011 (Flight 132), an uncommanded roll was attributed to a lateral-directional disturbance aggravated by a deactivated yaw damper. Gulfstream did not perform an in-depth aerodynamic analysis of these events before the accident. The Board concluded that if the company had done so, it could have recognized that the actual in-ground-effect stall AOA was lower than predicted.

The investigation also found that the target V2 speeds provided to the crew were erroneously low. Gulfstream based the G650's V2 speed on the minimum allowed by regulation (1.13 times the reference stall speed) but used speed increments from the G550 airplane to determine liftoff and rotation speeds. The Board concluded that this method was inappropriate and resulted in target V2 values that were too low to be achieved.

Rather than investigating the root cause of the V2 overshoots, the company focused on refining pilot technique. During a February 2011 flight, the team experimented with abrupt column pull forces of 70 to 75 pounds, which reduced the overshoots but raised concerns about whether FAA certification pilots would accept the technique. The difficulties were exacerbated in late March 2011 when the target pitch angle was reduced from 10 degrees to 9 degrees without an accompanying increase in takeoff speeds. The Board concluded that Gulfstream failed to properly validate the takeoff speeds using simulation or dynamic analysis.

The Board found that Gulfstream maintained an aggressive schedule to obtain FAA type certification within a 5-year window ending in September 2011. This schedule pressure was not adequately counterbalanced by robust organizational processes. The company's Flight Test Standard Practice Manual had not been updated in more than 12 years. As a result, individual engineers assumed multiple roles; FTE1 was responsible for test coordination, data analysis, report writing, and monitoring the flight control system. The Board concluded that deficiencies in technical planning, oversight, and safety management contributed to the incorrect speeds and the failure to adequately monitor critical safety parameters during the test.

### Probable cause

The National Transportation Safety Board determined that the probable cause of this accident was an aerodynamic stall and subsequent uncommanded roll during a one-engine-inoperative takeoff flight test, which were the result of (1) Gulfstream’s failure to properly develop and validate takeoff speeds for the flight tests and recognize and correct the V2 error during previous G650 flight tests, (2) the G650 flight test team’s persistent and increasingly aggressive attempts to achieve V2 speeds that were erroneously low, and (3) Gulfstream’s inadequate investigation of previous G650 uncommanded roll events, which indicated that the company’s estimated stall AOA while the airplane was in ground effect was too high.

Contributing to the accident was Gulfstream’s failure to effectively manage the G650 flight test program by pursuing an aggressive program schedule without ensuring that the roles and responsibilities of team members had been appropriately defined and implemented, engineering processes had received sufficient technical planning and oversight, potential hazards had been fully identified, and appropriate risk controls had been implemented and were functioning as intended.