X-31 Departure Flight Test Analysis and Support
Bihrle
supported the X-31 Enhanced Fighter Maneuverability (EFM) Program as
early as 1984, cooperating in development programs with Rockwell International,
Messerschmitt-Bolkow-Blohm GmbH, DARPA and NASA.
The aerodynamic capabilities of newer tactical aircraft, such as the X-31, have made modeling of the low-speed end of the envelope increasingly important, particularly at very high angles of attack. The X-31 planned to exploit this region during evaluations of tactical utility of the
experimental aircraft with demonstrations of high-angle-of-attack, post-stall, 180° turns, known as the Herbst maneuver.
During envelope expansion flight tests conducted by the International Test Organization (ITO) at NASA Dryden, the aircraft experienced a departure from controlled flight as the pilot was performing a maneuver at 60° angle of attack. Bihrle was tasked to analyze the flight data and
make recommendations as to the source of the departure and other unexplained post-stall behavior.
The X-31 aerodynamic model obtained by Bihrle from NASA Dryden was found to be inadequate in fully modeling the subsonic region, particularly at the high angles of attack. The model also did not contain adequate aerodynamic data to simulate the departure found in flight near 60° angle
of attack. On the basis of observations made by Bihrle it was felt that certain changes to the Dryden X-31 simulation database and data implementation techniques would improve the modeling of the in-control and departure characteristics of the flight test vehicle. This task specifically
focused on the simulation database requirements for these flight regimes and on the optimal method of incorporating forced-oscillation (body axis) and rotary-balance (velocity vector) dynamic data into the flight simulation database.
Comparison Of Model Response With Flight Test Results
A number of X-31 flights were investigated in this effort, including the departure of Flight 2-73. The maneuver was performed at 35,000 feet/0.4 Mach and consisted of a full aft pitch input from inverted flight with maximum afterburner set and the angle of attack limiter set at 60° (a
split S maneuver). As the aircraft approached 60° a positive yaw rate and negative sideslip excursion developed. The airplane continued to depart from controlled flight with increasing angle of attack and yaw rate. The angle of attack reached a value of 70°, where, the aircraft becomes
highly damped in yaw (a result found from previous tests conducted by Bihrle), preventing the aircraft from obtaining further increases in angle of attack resulting in a fast, flat spin. The pilot initiated recovery with forward stick and with the angle of attack reduction, the yaw rate
damped to zero, completing recovery of the aircraft to controlled flight.
The yawing moment propagation for the departure of Flight 2-73 compares the results of the original simulation model and the modified model response to the flight
extracted yawing moment. Also included are the contributions to the total yawing moment of the static offset, rotational effects, thrust vectoring and forced-oscillation damping terms (developed when the rotational effects are turned off). As shown in this figure, a result that was close to the airplane response only occurred when both the offset effects and rotary-balance terms were included.
Further evidence of the importance of modeling the rotary terms, particularly in the post-stall region, was found in the analysis of Flight 2-85, a 40° angle of attack 360° roll about the velocity vector. Initially, as shown in the time history traces, the trailing edge flaperons
followed the roll stick command, but then quickly deflected against the roll as dictated by the FCS software to control roll rate. However, even with the trailing edge flaps deflected against the roll the aircraft continued to roll in the original commanded direction, evidence of a
significantly propelling condition at this angle of attack.
The propagation of rolling moment versus time for the flight test data as well as the original simulation, which did not incorporate the rotational data showed that the original model exhibited a poor match with flight. Because the model, as originally mechanized, would
roll in the opposite direction when driven with the flight test control inputs, the test center attempted to improve the match by significantly reducing the lateral stability. Even with these changes, the match was less than satisfactory. Using the rotational data to define the damping
for this velocity vector maneuver resulted in a very good match with flight test data.
Summary
The original NASA Dryden X-31 simulation database exhibited significant divergence from the response of the aircraft seen in flight at many high angle-of-attack conditions. The work performed by Bihrle significantly enhanced the fidelity of the simulation model. Comparison of X-31
flight test results to the updated model response, including out-of-control motions, showed greatly improved correlation and led to a better understanding of the flight vehicle's behavior.
Other Flight Test Support Success Stories
Advanced Flight Test Support
F-16
MCID Configuration Development Flight Test Support
