Air Accident Investigation Unit report into serious incident Sept 23 2007
 

The AAIU have published Final Report No. 2008-004 on www.aaiu.ie concerning a B737 and MD-83
It's URL is <http://www.aaiu.ie/upload/general/10423-0.pdf
SYNOPSIS
Two passenger aircraft entered Irish controlled airspace near Reporting Point BANBA at high level off the South East coast of Ireland. One was an MD-83 routing northwards towards Dublin from Faro, and the second was a B737 routing westwards towards Cork from Stansted. The former was maintaining Flight Level (FL) 280, while the latter, which was cruising at FL300, was cleared initially by the Shannon based Air Traffic Control (ATC) Radar Controller to descend to FL290 and subsequently to FL100. This ATC clearance conflicted with the path of the northern bound aircraft, and, in spite of at least four warnings, one verbal and three electronic, the Radar Controller appeared not to comprehend the closing speeds of the two aircraft and allowed the higher one to descend and lose the required minimum vertical and lateral separation from the other. What ensued was a critical failure of the human element of the ATC system to rectify this situation. The last resort safety net in this extreme circumstance, each aircraft’s on board Traffic Alert and Collision Avoidance System (TCAS), automatically activated with a commanded warning to each aircraft. The pilot of each aircraft reacted correctly to this TCAS warning, one climbed his aircraft as instructed by the system and the other descended his aircraft as instructed by the system. A potential mid-air collision was thus narrowly avoided due to the TCAS activation and the correct response of the pilots. With separation subsequently re-established by ATC, both aircraft continued onwards and landed at their respective destinations.
The Air Accident Investigation Unit (AAIU) of the Department of Transport was notified of this Serious Incident shortly after it occurred. Three Safety Recommendations are made as a result of this Investigation.

Controller back on the job after training!!! Gulp!
 

More on this from the Irish Independant
http://www.independent.ie/national-n...r-1307958.html

New info high-lighted below.

AN air traffic controller who allowed two passenger planes to come within 17 seconds of a mid-air collision is back on the job despite an air accident investigation report concluding that the near disaster was the result of "human error".

The radar controller and a planning controller, who are both stationed at Shannon, were immediately relieved of their duties and had their licenses withdrawn following the incident on September 23 last year.

The radar controller, who had less than two years experience in the job and was responsible for guiding the two planes, underwent extra training and has since returned to work.

An Air Accident Investigation report published yesterday found that the safety of the two planes was "seriously compromised". They came within 600 feet of each other and within 17 seconds of impact.

Tragedy was only averted when emergency safety systems onboard the aircraft alerted the pilots and they took evasive action.

The two planes got so close that the pilot on the Flightline aircraft said afterwards that he had "brief visual contact" of the other plane.

The probe found that the radar controller received four warnings -- one verbal warning from his planning controller and three electronic warnings - but failed to react to them on time. Instead the Traffic Alert and Collision Avoidance System (TCAS) on board the aircraft kicked in and instructed the Ryanair pilot to climb and the Flightline pilot to descend.

The report found that the controller had "inexplicably" cleared the Ryanair Boeing 737 to descend towards the path of the other aircraft, an MD-83.

Air traffic in the sector was light to moderate at the time and the two controllers were eight minutes from the end of their nine and a half hour shift. They also had a rest break less than an hour earlier.

Investigator Frank Russell said both controllers were sufficiently aware both aurally and visually that the two planes were on a collision course. But while they were both individually aware of the problem, their collective response was inadequate to the task, he added.

The planning controller, who has 30 years experience, pointed out to his colleague that there was a substantial speed difference between the two aircraft, however he took no further action as he was engaged in other duties.

Mr Russell said the younger controller had a "partial loss of his situational awareness" which led him to take incorrect and potentially dangerous decisions. His belated attempts to rectify the situation were overtaken by the aircrafts' TCAS. "The safety (of the two aircraft) was seriously compromised by the failure of the radar controller to maintain a safe flight level between the two aircraft. There were no electronic/technical reasons which contributed to this failure. This was human error," concluded Mr Russell.

He said disaster was averted by the last line of defence, the aircrafts' on-board TCAS, and not by the intervention of air traffic controllers.

"Near miss in Ireland" Their words not mine!
 

Papers in Ireland reporting a near miss between two aircraft in Irish airspace last year. One a Ryanair 737 and the other an MD83. Journalistic dramatic reports of separation of only 150ft. Non dramatic report can be found here.

shoey1975
 

not seen the journalistic report, but the official one doesn't make for happy reading.
The report mentions min separation of 600ft and 3nm, compared to legal minima of 1000ft and 5nm. Just how serious is that? The very fact that TCAS appeared to save the day suggests it's no laughing matter, since it's often not to be 100% relied upon.

In other such airprox reports eg on approaches, the minima are a lot lower eg 3nm or even 2.5nm.
Could somebody explain why this is the case please? And is there a rough n ready guide to what the minima are in different airspaces?

Thanks -- any information received in strictest confidence if you so desire.
Ian Shoesmith
BBC News
[email protected]

What do you mean it's not to be 100% relied on?

When a TCAS RA (Resolution Advisory) is activated, it supercedes all ATC instructions, and it to be followed precisely.

Ian,

Of course, the answer is, "it depends."

My expertise is in the US system, so I can't speak for the UK, Irish, or any other minimums. However, in general terms it is a question of whether one or both of the aircraft are RADAR identified, what type of RADAR is in use and whether it is mosaic or not, how far the aircraft are from the antenna, and other factors (including wake turbulence and a question as to whether the aircraft are being separated based on RADAR or rather, visually)...

Here's an example of 2.5 mile separation from the US (JO7110.65S, 5-5-4. MINIMA...):

[snip]



g. TERMINAL. 2.5 nautical miles (NM) separation is authorized between aircraft established on the final approach course within 10 NM of the landing runway when operating in single sensor slant range mode and aircraft remains within 40 miles of the antenna and:


PS Also note: you said "...compared to legal minima of 1000ft and 5nm..." It's actually an OR. i.e. 1000 feet OR 5 nm...

PPS ICAO Annex 11 is also your friend here...

Shoey,

again, it depends. Standard sep is 1000ft/5 miles.

In approach control the CAA allow this to reduce to 3m/1000ft

for a tower controller such as myself, in theory separation can be as little as 'as long as they dont hit each other' in the form of 'reduced separation in the vicinity of the aerodrome' which is basically where they can get as close as they want as long as i can see them both/they can see each other and happy to maintain separation.

Below ( i think) 600ft (someone can correct me on this?) tcas RA can be ignored to allow this type of sep.

Also, IFR aircraft can be separated from VFR aircraft simply by giving each a/c 'traffic information' on each other (telling each where the other is) and they can also get as close as they want...

PositiveRate:

Think DHL B75F and TU154 over Swiss airspace a few years back.

Human error is always possible.

MM

Yes, but the issues with some nations pilots being trained to give ATC precedence over TCAS were cleared up after that event, so such a cause-effect chain is unlikely to happen again.

probably a good idea not to talk to the bbc either as they will only do their usual hatchet job and add the story to their "we hate aviation" editorial stance.

not interested in "hatchet jobs" as I hope this demonstrates: http://news.bbc.co.uk/1/hi/uk/6236810.stm
that followed about four months of work last year.

Anyway, I digress. The point I was making about certainly not wanting to rely on TCAS is surely responsible. Swiss Cheese anybody?

Follow the RA not the controller
 

The RA in that TU154 case was overridden by the crew choosing the controller command rather than the TCAS RA. There really is only one case in the vertical where you cannot obey, a GPW takes precedence. Flying at max altitude is the other case. Even there, a climb and stall is to be preferred over not taking the RA.

TCAS III would have allowed RA's in the horiz plane as well, and would be a good alternative to the GPWRS precedence that we have today with vertical RA's

Published Report
 

Dear Ian,

Some how I find it hard to belive that it took four months to gather all that information. If you were to look at the CAA and BALPA sites and ask it would take probably a few hours at most to gather the statistics that you and your partner had.

An autopilot system does not just begin to 'turn the aircraft' it will follow the planned route or it must have an input to the system by someone. With all the available information on the World Wide Web surely you were able to find that out.

There are many factors to all situations and incidents, the 'Swiss cheese effect' that you already mentioned, the TCAS system is another line in defence to avoiding traffic conflicts. It will 'talk' to the other TCAS encoder and they will work out a 'resolution' to the situation.

As far as i am concerned they should be followed unless that puts you in even more danger but that would hopefully never happen, wont say never but...

I do hope that you will lead the way of you and your fellow journalists by reporting true and factual pieces and not misinterpreting or misrepresenting the information provided by persons on this site.

all the best

259

A disappointingly patronising and poorly-judged comment.

I may be wrong, however I believe TCAS RA's are inhibited below 1000 ft, (it may be 500ft) and the TCAS will revert to TA only below this.

While this may not excuse this incident, the following is at least an indication of the level of service offered
TCAS when operated by a trained crew (ie the guys and girls who follow it) is a highly effective safety system. The worry comes from the possibility of Very Light Jets (VLJs) mixing it up in both controlled and uncontrolled airspace without TCAS fitted, as discussed here...
http://www.pprune.org/forums/showthread.php?t=312649

Of course Ian Shoesmith's comment that one does not want to rely exclusively on TCAS is both responsible and appropriate.

I continue to be surprised at how often TCAS and people's opinions on its trustworthiness or not come up in this forum. It seems that there are plenty of people who want to say that of *course* TCAS is the best thing since sliced bread, but some must also be aware that there are weaknesses, otherwise it wouldn't keep coming up.

TCAS is a sociotechnical system that not only has its weakness in that the human part does not consist of perfectly reliable robots but also inter alia that the algorithms are not yet known to be adequate for other than two-aircraft conflict (and as far as I can tell might be inadequate for three), and in that it can present pilots with decision problems which do not have an adequate solution. If you talk to ATCOs who work on TCAS issues you will also learn of more issues.

I am afraid that 757 Driver's wish that remains that, a wish. Things have gotten better, but they are by no means uniform. All kinds of different organisations still give somewhat differing guidance on actions in response to an RA.

Let me direct people towards the Tech Log thread TCAS Philosophies for discussion.

Those wishing for some engineering-scientific detail may care to look at my paper on Causal Analysis of the ACAS/TCAS Sociotechnical System, an invited paper in a safety-critical systems conference in Brisbane in 2004, and the paper on TCAS at the same conference by Ed Williams, Airborne Collision Avoidance System

The crew did follow the controller's descent advisory, but we don't in fact know what their reasoning was. The commander was concerned that they were involved in a three-aircraft conflict and was searching for another aircraft that they did not see, and which he was concerned was the traffic in the controller's advisory.

Apart from the fact that these two sentences contradict each other, I should point out that you don't in fact know that. The complete situation with three-aircraft conflicts remains unknown. For some three-aircraft conflicts, the TCAS algorithms will suffice to avoid collision, but it is not known for all.

PBL

In the hope of putting out some anti-journo flames:

Mr Shoesmith said:
"The very fact that TCAS appeared to save the day suggests it's no laughing matter, since it's often not to be 100% relied upon."

I took him to mean that TCAS is not infallible, and that you should not rely on TCAS to avoid collisions. If you get an alert, something has already gone wrong.
I did NOT take him to be saying that pilots should not follow TCAS directions, if the thing goes off.

I can see the potential for ambiguity, but "not to be relied on" is not the same as "not to be obeyed."

Peace and love

One of the objectives of ATC is to prevent collisions between aircraft. They acheive that objective through a number of means many of which do not involve radar.

When using radar the objective is still simply to prevent the aircraft from being in exactly the same place at the same time.

Radar is not 100% accurate.

Depending on the radar performance - scan rate, frequency, pulse length, PRF clutter reduction elements etc, the number of radars used, the range at which the aircraft is from the radar among many other considerations lead to a minimum horizontal separation being set that will ensure aircraft which are at that minimum separation will not collide.

This separation can be 10nm, 5nm, 3nm or in special cases some figure less than 3nm.

It must be remembered that when two aircraft are displayed on a controler's screen with the centres of the blips or radar position symbols exactly the minimum appart, the aircraft can actually be less than that distance appart however the system guarantees that they will not be so close as to be in danger of collision.

Thus when at large range from certain radars, 5nm indicated separation may not be suficient and therefore 10nm will be prescribed in those cases.

-----------

With regard to TCAS, it must be remembered that the TCAS system scans for other aircraft which will enter a small zone arround the parent aircraft. If it predicts that another aircraft will enter this zone it issues an alert to the parent aircraft's pilot to either climb or descend to prevent the other aircraft from entering that small zone.

It is important to note that again the TCAS system is very accurate but not 100% so. Therefore since it is important to err on the side of safety in many cases it issues a warning when aircraft will get very (dangerously) close but when they will not actually collide. i.e. they will actually miss by some metres or even 10s of meters.

Thus when reporting this serious incident report, while not detracting from the importance of the report and the serious nature of the incident, it is impossible from the radar or from the TCAS to say 100% that the aircraft were on a collision course. They would have got dangerously close no doubt and TCAS ensured a minimum level of safety but no one can say in certainty that they would have collided.

Many people in the industry will be familiar with the ATC advice of traffic opposite direction straight ahead 1000ft above report when clear and on the ATC radar the indications of the relative positions close and merge to the same position before separating again but the pilots see the aircraft pass a small distance to the left or the right of their aircraft.

So controllers try to prevent colisions. They try to use various techniques to keep aircraft a certain defined indicated distance appart so that they do not collide. TCAS works to keep a small safe zone round the parent aircraft clear of other aircraft who while very close may not be about to collide with the parent aircraft.

Regards,

DFC

PS The report did not dwell on one very important issue. With the display on 250nm range measure the distance on the screen between two tracks 10nm apart. Now change the range to 150nm. Those aircraft appear to move appart. With a tired mind, end of shift distractions, cluttered airspace lines etc it would be easy to allow one's mind to use the same visual on screen distance to represent a minimum of 5nm when now it is actually less than 5nm.

STCA suffers from false indications especially when aircraft are changing level at high rate.

Combine the two and one could see a situation where the controller could think that they had more distance than they actually have and the initial STCA would be initially counted of as unreliable.........but then on checking the controller would realise the error and take corrective action - as in this case.

Is there a 5nm reference available to the controller in that area of the sector?

References to the South German accident may not be relevant since I believe that maintenance was on going and the STCA was not available in that case.

A very constructive post by dfc and a possible insight overlooked by the aaiu team which he mentions in his ps comments. At a 250 mile range the physical distance of 5 miles appears much "closer" on a radar screen than 5 miles looks on a radar screen at 135 mile range.

There is no 5 mile reference anywhere in the system. One can be drawn by using the qdm tool and some do just to remind themselves what 5 miles "looks like" when operating at a much smaller range after operating a period of operating at much bigger ranges.

How many Atco out there in other centers operate on ranges out to 250 miles regularly?

One thing of interest to me is that when the initial instruction was issued for the ryr to descend thru the flt a third non involved aircraft (malev) i think took the call. .This necessated the controller to make another transmission to tell the malev to disregard the intruction and then reconfirm the instruction with the ryr-which lost a valuable amount of time . I just wonder if the ryr had gotten the first call would it all have worked out ok? Who knows.

Some more credance should also be given to the "sytemic,training and procedural" issues identifed in the report. Human error is given as the primary cause but the other factors are just as important. TRM every 3 years now recommended to once a year/year and a half. The TRM itself to be changed and include a culture of active "challenge" as pilots do.
That a study be performed and recomendations made on the on-suite allocation of tasks between the radar and planner.And a standards officer to be appointed to co-ordinate between operations and training. . Of the 3 recommendations all 3 three are important but emphasis should be given to the first 2.

On a purely personal note-a message of support and good will to those involved. It cant be easy but your colleagues support you.

Yours as ever
OCK1F