Unreliable airspeed events pose a significant challenge (and safety risk) because such situations throw onto aircrew the most difficult (and error prone) of human cognitive tasks, that of ‘understanding’ a novel situation. This results in a double whammy for unreliable airspeed incidents. That is the likelihood of an error in ‘understanding’ is far greater than any other error type, and having made that sort of error it’s highly likely that it’s going to be a fatal one.
Then of course there’s the nature of air data sensors themselves, while pneumatic sensor heads may be carefully mapped into aircraft ‘swim lane’ style processing channels, in practice they still share common plumbing, externalities and working principles. So should we be surprised by a string of common cause failures of such systems? Probably not. All this because we still rely on sensors whose basic working principles would be understood by the Wright brothers, hell even by Leonardo Da Vinci.
My question is, can we can do (a little) bit better?
As luck would have it over in the fighter aircraft community the problems and limitations of air data sensors are well understood and alternatives to pneumatic sensing of air data has been a busy research area for many years. One of the incentives is that in air combat there’s a distinct tactical advantage to being able to fly extreme ‘high alpha’ manoeuvres, but in doing so you can (as an example) also run smack into the mechanical limits of the angle of attack sensor vanes. Given that military aircraft can have a tendency to depart abruptly during these sort of situations (1) loss of a critical input to your flight controls like AoA is not as they say, a ‘good thing’.
A practical example to illustrate. On the F/A-18 aircraft at alpha angles above the mechanical limits of the AoA probes an estimator is used to generate an AoA value (Marshall 2004). The estimator uses actual stabilator position and a ‘look up table’ approach, with estimation a function of aircraft gross weight, Nz, and Cz (normal force) (2). During flight within the normal range of the AoA probe the estimator value is blended with the sensor values with the weighting progressively skewing towards the estimator as alpha approaches the limits of angular measure of the traditional vane sensor.
The F/A-18 example is interesting for a number of reasons, first that it’s been fielded into an operational fleet (the F/A-18 D/E/F) so it’s not a theoretical solution but one in service. Secondly the objective of the F/A-18 program was to implement a safety of flight improvement with minimal impact. As a result it uses the existing flight data available within the flight control computers, and there are no hardware modifications. Thirdly the estimator was not used to supplant the existing AoA probes or air data functions (imagine the certification hurdles for that) but instead provide a diverse source that could be fused with the hardware sensors in parts of the flight envelope where their data became questionable (3). Finally a diverse AoA source allows hardware sensors that have been excluded due to hitting the stops to re-enter their respective computational channel on the basis of matching the estimator.
Adding such a diverse source of air data would go a long way towards moving civilian aircraft’s air data systems from their current position of robust fragility to one of resilience. Just as obviously the adoption of a flight proven solution that requires no additional hardware has practical advantages in terms of rolling out such a modification in the real world of customers, dollars and regulators.
Mitchell, E.J., “F/A-18A-D Flight Control Computer OFP Versions 10.6.1 and 10.7 Developmental Flight Testing: Out-of-Controlled Flight Test Program Yields Reduced Falling Leaf Departure Susceptibility and Enhanced Aircraft Maneuverability, Master’s Thesis, University of Tennessee, 2004.
Zeis, K.E., Angle of Attack and Sideslip Estimation Using an Inertial Reference Platform, AFIT/GAE/AA/88J-2, Masters Thesis, Air University, 1988.
1. See for example the falling leaf spin mode in the F/A-18 aircraft, which has claimed a number of aircraft and pilots over the years.
2. The Alpha estimator is based on work done for an F-15 AoA and sideslip estimator using an inertial reference and a Kalman estimator (Zeis 1988).
3. Sideslip is likewise calculated.