QF72: Did a cosmic ray zap the Airbus?
The Australian Transport Safety Bureau is now considering the remote possibility that a rogue cosmic ray or solar particle caused a Qantas A330-300 to twice dive out of control over Western Australia on October 7 last year.
Startling though this may sound, the second interim report into the accident, in which 12 people were seriously injured and at least another 107 suffered minor injuries on the flight from Singapore to Perth has now all but eliminated, as a factor, artificial electromagnetic interference from personal computers, the jet’s own computers, its inflight entertainment system or even the powerful military transmitters at the Harold E Holt naval communications base.
But was natural high energy particle damage a factor? That is the new question.
The ATSB is taking seriously every possible factor in an accident which has defied a complete explanation despite a rigorous examination of the jet’s systems, their maintenance, and everything else that occurred without warning that day, forcing an emergency landing at the Learmonth base, not far from the Holt transmitters.
The second interim report confirms that for whatever reason one of the three air data inertial reference system units or ADIRUs which inform the flight control system of the jet about its attitude or angle of attack among other things was able to overwhelm its error protection system with spurious data.
This set in train, very abruptly, two violent dives generating the sorts of positive and negative G forces that most people other than top gun military pilots would only experience in amusement park thrill rides.
The ATSB reports (both need to be read carefully) detail the exceptional challenges the pilots had to overcome to regain control of the jet, and the confused state of the electronic error messages that were generated as they headed for Learmonth.
But nothing has been found in the manufacture or maintenance or operation of the US made ADIRUs or indeed any other mechanical or systems related function to explain why things went so wrong.
Airbus has however since changed the filter rates and other processes in the ADIRU and related systems to make it ‘highly unlikely’ that such spurious spikes in data can ever again cause a similar upset.
Which leaves the unlikely, but troubling question about high energy or solar particles hanging in the air. To quote the report:
There is a constant stream of high-energy galactic and solar radiation interacting with the Earth’s upper atmosphere. This interaction creates a cascade of secondary particles. Some of the secondary particles, in particular neutrons, can affect aircraft avionic systems.
A single event effect (SEE) can be:
- non-destructive, involving a soft error, where a logic state in a digital electronics component changes from a 1 to a 0 or vice versa but can be reset by cycling the power off and on;
- or a hard error,resulting in permanent damage of a component that is not recoverable, even by cycling the power off and on.
High density integrated circuits, such as memory devices and central processing units (CPUs), can be particularly susceptible to SEEs. SEEs have been suspected of generating some of the soft errors that occur in a wide range of different aircraft systems. Hardware and software design features such as redundancy, monitoring, error correction and partitioning can be useful to mitigate the effects of SEEs.
The investigation team is evaluating the relevance, if any, of SEEs to the ADIRU fault that resulted in spikes being produced in ADIRU parameters.
The ATSB report also details the dissimilarities between the QF72 incident and the disaster that killed all 228 people on board Air France flight AF447 on June 1. (My view is that problems with the weather radar on that flight may have been at play and combined with the iced up pitot problem already discussed by the incomplete French investigation converged on the crew of that flight, in an Airbus A330-200 with tragic consequences.)
There is a risk the ATSB’s raising the issue of high energy particle damage may be confused with the risk to aircraft, and a whole range of computer and power grid reliant processes in everyday life at large, posed by a several types of severe solar storms. These are different from SEEs in that a network of satellites and earth based telescopes will provide timely warning of such a large scale event, similar to those that famously occurred world wide in 1859, or blacked out Quebec in 1989.
Categories: Uncategorized
You must be logged in to post a comment.
A somewhat tangential question, but do you know how much, if any, truth there is in the concept that cabin crew and pilots (particularly those who work for many years on International routes) have a significantly higher incidence of cancer due to the higher exposure to cosmic rays? (since at altitude there is less shielding from the atmosphere).
by Bogdanovist on Nov 18, 2009 at 7:56 pm
Haven’t seen any current studies. In the late 90s SAS was reported as limiting exposure of female cabin attendants to flights passing near the magnetic pole or higher because of medical advice that the atmospheric flux was elevated in those regions where particles could be pulled together more tightly by the shape of the planet’s magnetic field.
In 2003 at the end of the Concorde era study of dosimeters worn by its pilots was found to have shown no evidence of effects of increased exposure to background radiation at higher than normal altitudes, but this may have also reflected over all shorter flying hours, as Concorde passed through higher altitudes much faster than subsonic jets at lower altitudes, skewing the relative exposure.
It has also been argued that the life style of pilots and flight attendants masked the true results, in that many of them were able to take (in the good old days) duty rest periods in sunny places and expose themselves to the risk of skin cancer more intensively than the general population.
I believe however that a study of Lufthansa pilots inferred in the 80s that those who crossed many times zones in a short period with high frequency across the North Atlantic exhibited shortened life spans.
Many years ago an acquaintance who was an Australian nuclear physicist of some note died in his early 60s of tumours that some suggested were the sacrifice for working in laboratories with higher than normal levels of radioactivity. However as his life was drawing to an early end he told his associates that he had exceeded the lower statistical limits for heavy smokers, lovers of fine red wine and sunbaking, all ‘vices’ that he enjoyed with enthusiasm, so his shortened life span proved nothing. Similar life style choices by any group of professionals would make a statistical analysis very difficult.
by Ben Sandilands on Nov 18, 2009 at 8:31 pm
Thanks Ben, I can imagine population studies like this would be a nightmare given the many and varied factors. I guess the take home message is if there is a danger from flying, it is a relatively small increase otherwise it would have been easier to spot even given the other factors.
by Bogdanovist on Nov 18, 2009 at 8:53 pm
I really have trouble believing that the autopilots in these fly by wire machines are so error-intolerant:
1. The autopilot disconnects when there is a spurious input. This is a typical engineer fix, but shows a lack of understanding for how the real-world operates. Just when you may need the box the most (e.g. iced-up pitots; EM interference) it throws up its transistors and disconnects! I would think that a default to a “last known good (straight/level) configuration” would be preferable to a disconnect. Even if the thrust settings are incorrect, wings-level will save lives.
2. Spikes are interpreted by the system as valid. How did this fault sneak through? Once again, bad/lazy programming of the device.
If you’re going to confer total aircraft control to electronic devices, they should be not only double-redundant (2 backups), but they should be equipped with worst-case default settings that don’t simply relinquish all control and leave the aircraft to be at the mercy of a possibly disoriented, frightened or incapacitated crew.
Cheers!
by TheGoodSkeptic on Nov 19, 2009 at 8:33 am
I can believe a cosmic particle causing SINGLE upset in a SINGLE IMU
But no way causing an upset in all the IMU’s.. No way… Just (statistically) not going to happen….
And as for EM interference causing the problem- anyone who even puts up such a possibility really doesnt understand the world of radio physics.
Glen
by glen@pacificmedia.com.au on Nov 19, 2009 at 1:31 pm
I can well believe that bored adolescent aliens take pot shots at aircraft, with their cosmic ray guns, just for fun.
by Mr Pastry on Nov 20, 2009 at 8:44 am
Nice to know there’s a future for good old fashioned hydro-mechanical airplanes — keep ‘em around for the next forecast Solar Max year…
by LongTimeObserver on Nov 21, 2009 at 5:56 am