View Full Version : European airspace safety rules 'flawed'?

14th Nov 2003, 04:19
As reported in the current edition of the New Scientist

European airspace safety rules 'flawed'

19:00 12 November 03

A leading air safety expert is warning of a serious flaw in the risk analysis undertaken before rules were changed to halve the vertical distance between planes flying over Europe. He says this means the risks of flying in European airspace are not as low as the public is being led to believe.

The warning comes from Peter Ladkin, a computer scientist specialising in dependable systems at the University of Bielefeld in Germany. He says that Eurocontrol, the pan-European air traffic control organisation, failed to take into account a clutch of air incidents and accidents involving sharp changes of altitude. This, he says, led to an unduly optimistic estimate of the risk when aircraft fly closer together.

The new flight rules, called the Reduced Vertical Separation Minima (RVSM), were introduced on 1 January 2002. By cutting the vertical separation between cruising airliners from 2000 feet to 1000 feet, the change doubled the number of aircraft able to fly in the busy cruising air lanes between 29,000 and 41,000 feet.

The ability of modern aircraft to maintain their height precisely means the new rules are safe, says Joe Sultana, who heads the RVSM project in Brussels. But Ladkin says the safety argument failed to take into account other factors that influence an aircraft's altitude.

Unpredictable response

In particular, he points to the unpredictability of the response by pilots and air traffic controllers to warnings from ACAS, the airborne collision avoidance systems.
ACAS uses radar on board the plane to look for approaching aircraft, and advises pilots to dive or climb out of trouble. But the reduced vertical distance between air lanes means that this could put the plane on collision course with a third aircraft much sooner than in the past.

Eurocontrol's safety case claims the reduced separation will lead to only one collision every 150 years. "But some recent air incidents show that might not be the case," Ladkin says.
Ladkin points to three cases - one mid-air collision in which 71 people died and two near misses - which expose the potential risks associated with RVSM, two of which involved ACAS.

Ladkin believes that more work is needed to study the way pilots respond to emergency instructions from ACAS. "I don't think we yet understand the behavioural side of ACAS interactions well enough to guarantee safety under RVSM," he says.

Updated analysis

Eurocontrol strongly disputes Ladkin's analysis, insisting that ACAS has no bearing on the safety of RVSM.
Using the first 18 months of flight data under the new system, Eurocontrol is updating its analysis of RVSM safety. It expects to publish the result in the next three months, but seems unlikely to change its assessment of the new system's safety.
"We have found that ACAS in RVSM has not caused any significant difficulties," Sultana says.
Ladkin is unconvinced by Eurocontrol's claim that the data will justify its claim that RVSM will lead to no more than 2.5 x 10-9 collisions per aircraft flight-hour or one collision every 150 years. To attain that level of confidence, he says, would take at least decades of data.

Pax Vobiscum
14th Nov 2003, 16:17
Links to the original papers can be found at Peter B. Ladkin's (http://www.rvs.uni-bielefeld.de/publications/Reports/SCflawed.html) home pages. There is also a discussion of the South German Midair (http://www.rvs.uni-bielefeld.de/publications/Reports/ACAS-Midair-www.html).

ATC Watcher
15th Nov 2003, 06:06
Thanks Pedota , very interesting post .

Prof Ladkin is a very controversial figure, but worth listening to. He is loved by Airbus, and now certainly by Eurocontrol.
I remember his work on the early A320 logic faults. Very impressive.

For the RVSM in Europe the Eurocontrol study indeed meets the Target Level of Safety . But is is based on total compliance with ACAS RAs ( i.e deviations of a few hundred feet max ) reality is indeed different and pilots reactions to RA are sometimes interesting to watch.
The point is that too little real data exists on the relations ACAS-RVSM to make a generality.

At the moment both Jo Sultana, the Eurocontrol RVSM Project Manager and Prof Ladkin are right. The RVSM TLS is indeed mathematically met , but is one RVSM induced collision per 150 years a realistic figure in real life ?