Italian investigators’ final report on the fatal 2015 AW609 prototype in-flight breakup highlights disparities between engineering simulations used to design the fly-by-wire civil tiltrotor and its actual performance in the air.
The May 10 report by the National Agency for Flight Safety, known by the Italian abbreviation ANSV, comes as the U.S. National Transportation Safety Board is wrapping up the factual portion of its investigation into the July 6, 2016, crash of another experimental fly-by-wire aircraft: Bell Helicopter’s 525. Airframe and flight-control vibrations preceded the 525 prototype’s in-flight breakup over Italy, Texas.
As they begin the analysis phase of their probe, investigators will want to understand the apparent absence of such reparations and previous 525 flight tests and the fidelity of extensive engineering simulations in predicting such airborne occurrences.
The ANSV concluded the No. 2 AW609 type broke up over Santhià, Italy, after the blades of its right- and left-hand prop rotors all flapped extremely and struck the wings’ leading edges during the test at the aircraft's design dive speed of 293 kt (calibrated airspeed). The blade strikes severed hydraulic lines, triggered a fire and led to the in-flight breakup that killed test pilots Herb Moran and Pietro Venanzi. The pilots already had performed dives that day of 303 and 295 kt.
The report presented the following account. During the fatal test, Moran was flying and had just completed a 180-deg turn to the left when the aircraft began experiencing oscillations in roll and then in yaw. Initially, the pilots did not counteract the oscillations.
The ANSV noted that such a roll and yaw had been experienced in previous flight tests and was considered self-damping phenomena. (The report also notes that the oscillations detected in telemetry data may have been too slight initially out to detect.) But the oscillations increased.
“Man, roll, and yaw!” Moran said about 15 seconds after the oscillations began, according to ANSV’s transcription of the cockpit voice recorder. He then began control inputs to counter the roll, followed by pedal inputs to counter the yaw.
However, the ANSV says, control inputs that deflect the AW609’s flaperons in roll swirls air toward the tail that induces yaw. The AW609’s flight control system is set up to anticipate that and send "feed forward" to the vertical stabilizer counter the yaw.
“Consequently,” the ANSV notes, “giving a command in counter-phase on the roll axis to dampen the relative oscillations creates an effect on the yaw axis that can be in phase with the yaw oscillations.” This happened during the accident sequence. Moran was manipulating flight controls to dampen the oscillations, but “the correction of the roll oscillation induced, by the control laws of the [flight control system], a maneuver in phase with the oscillations on the yaw axis.” That amplified the oscillations.
As a result, the aircraft entered a severe sideslip of 10.5 deg. (The design limit for that flight condition is 4 deg.) That led to extreme blade flapping and strikes with the leading edges. The aircraft's breakup followed shortly thereafter.
ANSV investigators traveled to AgustaWestland Philadelphia Corp. to attempt to replicate the accident sequence’s flight conditions in the companies SimRX engineering simulator. The simulator had been updated to reflect AW609 configuration changes to incorporate a more streamlined, tapered aft fuselage and a reduced vertical stabilizer surface.
“With the SimRX configured in the same software and flight conditions of the accident, however, it was not possible to reproduce the conditions occurred during the accident,” the ANSV says. The simulator’s inability “to faithfully reproduce the dynamics that occurred during test flight” may have been due in part to an aerodynamic data set that did not represent the extreme conditions of the final flight.
Among the safety recommendations included in report, the ANSV calls on the FAA and the European Aviation Safety Agency “to verify that the aerodynamic behavior of the aircraft at high-speed conditions will be reviewed, if necessary making use of wind tunnels tests in addition to updated models and simulations that can be representative of the complex flight conditions of this peculiar aircraft.”
The ANSV also called on those agencies “to verify that the control laws of the aircraft will be reviewed in the management of the extreme flight conditions in which the aircraft could possibly fly. That verification should be addressed to ensure the effectiveness of the flight controls inputs given by the pilot avoiding the possibility of unexpected and un-commanded coupling effects.”