TCAS philosophies

[joernstu]

All of a sudden I get a TA with the TCAS display showing traffic at my altitude and converging from the left. Moments later ATC says traffic at two o'clock, descend immediately and nearly simultaneously, I get a RA saying to climb. There is no other traffic displayed on my TCAS display. I do the rational thing and follow the RA and everyone survives this conflict.

This rational action did not prevent the DHL Boeing from colliding.

For the Tupolev Crew: You forget, that they communicated with ATC for several seconds before they got the TCAS RA. They initiated the descend before TCAS issued its RA.

Your whole argumentation on rational action is biased by your western training. The Tupolev crew wasn't so "furtunate". In Russia the controlling of traffic is handled differently if you are an international flight that if you are a national flight. For national flights the controller has the authority to command the crew to take a certain action.

I know.......Maybe there is some military flight out there that has an emergency which just happens to be affecting his transponder at the very exact time that I am getting an RA. I'm sure there are all kinds of obscure scenarios that can be thought of. Conspiracy theorists do it all the time.

If you have followed this threat, you will see that a Ueberligen-like mid-air is still possible today, it sounds extremely simplicistic to present insufficiencies in the TCAS system as "obscure scenarios" of "conspiracy theorists".

[CDN_ATC]

What about this:


747 off of JFK full of fuel taking off into the NAT tracks, operating at maximum available cruising altitude for the moment at FL360, ATC asks if they're able FL 370 for traffic earlier, they are not able that level.

Now my question IF TCAS asked them to climb for an RA, is the airplane going to get up there? Also does TCAS take that into account or is it completely separate from those systems?

Since airplanes flying long haul are often operating near or at their maximum cruising altitude for their current weight, what if the RA tells them to climb and they cannot? Then what?

[PBL]

Originally Posted by punkalouver

All of a sudden I get a TA with the TCAS display showing traffic at my altitude and converging from the left. Moments later ATC says traffic at two o'clock, descend immediately and nearly simultaneously, I get a RA saying to climb. There is no other traffic displayed on my TCAS display. I do the rational thing and follow the RA .........
By the way, if in this scenario there actually was an intruder at my two o'clock with no altitude given as pointed out by ATC(perhaps due to mode C not working), I would not go opposite to the RA.

This is wonderful. You give us two different courses of action depending on whether there isn't an airplane where ATC says it is or whether there is.

Need I say it: if you have this trustworthy information, then you don't need either TCAS or ATC.

Let me show you a different way of thinking (I emphasise the word "thinking" ).

You paint an intruder at 10 o'clock. ATC gives you a descent for traffic at 2 o'clock. You don't descend (which would be against the RA).

BAM! In the few seconds of consciousness available to you, you might like to consider where your reasoning went wrong. Maybe it was because the likelihood is rather high that there is actually an airplane where ATC implies there is an airplane, even if your on-board kit doesn't see it; or maybe .... sorry, that's all folks.

I'll offer to write your epitaph. First proposal: "I've proved I'd do what I said I'd do".

PBL

[PBL]

Originally Posted by CDN ATC

what if the RA tells them to climb and they cannot? Then what?

Then they don't climb, and their "dance partner" is issued a Reverse RA.

At least, that is the thinking behind Reverse RAs. That it doesn't always work that way was proved at Überlingen.

PBL

[punkalouver]

Originally posted by joernstu
This rational action did not prevent the DHL Boeing from colliding.

This rational action(FOLLOWING THE RA) would have prevented the collision as the report obviously states.

Originally posted by PBL
This is wonderful. You give us two different courses of action depending on whether there isn't an airplane where ATC says it is or whether there is.

I don't believe so. Read my previous post completely and carefully this time. In both scenarios I am not going to go against the RA's(as I have continuously stated on this thread) whether there is traffic at two o'clock or not. That is what I meant and that is what I said

You already have 71 epitaphs to write. No need to write my epitaph or anyone else's in this scenario because I would have FOLLOWED THE RA and ignored ATC once it was contradicted by the RA and I will do so in the future.

Why, because I have done my "thinking" before the flight even started and realized that Following the RA is the safest thing to do in a scenario like this, not trying to figure out the situation with little time available.

It is very true that the Russian system of training was at least partially at fault in this accident as has been mentioned in the report, however I was asked basically what I would do in this situation. Or maybe the captain listened to a couple of guys on an web site somewhere questioning the prudence of FOLLOWING THE RA.

More example of people who don't follow RA's or who try to out think a situation.

http://www.arinc.com/tcas/acasll_bulletins.html

Originally posted by CDN_ATC
Since airplanes flying long haul are often operating near or at their maximum cruising altitude for their current weight, what if the RA tells them to climb and they cannot? Then what?

There are situations where an aircraft is unable to comply fully with an RA due to performance limitations. The TCAS can be configured(not by the pilots) so that no climb RA will be given above a certain altitude for your aircraft which may or may not be your maximum certified altitude. When at a max altitude due to weight or in a low energy state or an non-normal configuration(gear/flaps that can't be retracted, etc) reacting fully to an RA can lead to stall warning. Use great care in such a situation to prevent an aircraft upset.

[CDN_ATC]

Thanks for those replies on my questions guys, really appreciate it.

I still find TCAS somewhat an imperfect system, if it can issue RA's when there's not an airplane near them, then I wonder if it will always issue the "right" direction.

I don't think the procedures should change, but something that must be taken into effect (that it's far from perfect)

[bsieker]

Originally Posted by punkalouver

[...] I was asked basically what I would do in this situation. Or maybe the captain listened to a couple of guys on an web site somewhere questioning the prudence of following an RA.

We already knew that you would always follow an RA.

You were asked to show the error in the argument that the Baskirian crew's actions were rational, and how you arrived at your conclusion that it was not rational.


Bernd

[punkalouver]

Originally posted by bsieker
You were asked to show the error in the argument that the Baskirian crew's actions were rational, and how you arrived at your conclusion that it was not rational.

Aside from the fact that they are all dead? Well, if you are flying in the mountains and ATC says descend immediately for traffic and moments later your GPWS goes off, which takes priority? If you are climbing slowly because you are heavy and ATC direct an immediate increase in climb due to conflicting traffic and moments later the stick shaker sounds, which takes priority? If you are in the situation over Germany that night and you have the conflicting messages that this crew had, which takes priority? There is only one rational answer to each, and in my opinion and none of them involve ATC taking priority.

But.....a previous post said that my action is biased by western training and that in Russia...."For national flights the controller has the authority to command the crew to take a certain action." It is a regulation to follow ATC instructions where I fly as well. Does that mean I am going to fly into the ground or stall the aircraft to follow an ATC instuction for immediate traffic separation? Do the Russian crews need to be told to ignore their stall warning or continue descending into the mountain peaks against their TAWS warning because of an ATC instruction?

P.S. I forgot to mention to the ATC poster that it is common to turn our transponder(TCAS) to TA only after an engine failure due to performance limitations. Also on parallel approaches to runways with less than 2500 foot spacing.

[joernstu]

"This rational action did not prevent the DHL Boeing from colliding."

This rational action(following the RA) would have prevented the collision as the report obviously states.

What would have been and what was are two completely different things. Fact is that the Boeing crew did behave just like you said you would. You said, that this would be the rational action. This must mean: following the rational course of action did not prevent the collision for the Boeing crew in the Ueberlingen case. The mid-air itself proves this.

Arguing about what-if is a completely different matter.

Had both aircraft followed the RA, the mid-air would have been prevented.
Had there only been the resolution issued by the Controller, the mid-air would have been prevented.
Had TCAS issued a reversal RA for the Boeing and Tupolev, the mid-air might have been prevented.
Had the STCA been available for the controller at ACC Zurich, the mid-air would have been prevented.

I could go on like this for pages ...

[punkalouver]

Originally posted by joernstu
What would have been and what was are two completely different things. Fact is that the Boeing crew did behave just like you said you would. You said, that this would be the rational action. This must mean: following the rational course of action did not prevent the collision for the Boeing crew in the Ueberlingen case. The mid-air itself proves this. I could go on like this for pages ...

It would appear that you are going on. In the end, the vast majority of us follow the rational route and accidents are prevented. A few don't follow the rational route either intentionally or by mistake and only luck or chance prevents an accident as in the JAL case which I provided a link to and I'm sure you read. Most properly functioning systems and societies depend on the vast majority of us acting responsibly. Of course accidents and deaths occur frequently because of those who break the rules and occasionally to those who don't.

I hardly think that because the Boeing crew doing the proper procedure and dying because of it should mean that as a result we should all just stop following RA's and do something different on the off chance that the other guy may be not doing what should be done. After all that would be irrational(or lacking in LOGIC). As I said earlier it is an odds game with a much better chance of survival happening if you FOLLOW THE RA.

And fortunately people who behave like this(ignoring the RA) are few and far between, although there appear to be three on this thread. Strangely(according to their profiles), they all appear to be from Germany, two from near Bielefeld(the third an unknown German location) and two are aircraft accident analysts. One even has logic as an interest. What are the odds of that?

[PBL]

punkalouver,

bsieker and joernstu and I work together. They have put a considerable effort into analysing the Überlingen collision, amongst other accidents, so you are discussing with people who understand these situations very well.

Your recent postings support my earlier suggestion that you have not understood the decision problem. Here is the reason.

I suggested a consequence of your point of view, namely that you could be involved in a collision. You replied in a way that suggested that you think it is not possible under the conditions described (at least, that is what you say).

Well, such a view is obviously mistaken. It is obviously possible that you could collide in such a situation, even to people to whom TCAS is the best thing since sliced bread.

Let me attribute to you some insight, and assume you agree that it is possible that you might collide. Then you must think it is extremely unlikely. So in that case I would ask you to show your reasoning for this unlikelihood.

What you have said so far shows that you think it unlikely that you would ever come in such a situation. OK, but you were asked to judge, not the absolute probability of being in that situation, but the conditional probability *given the situation described*, for that is the judgement on which a rational decision is based. And the conditional likelihood can be a very different quantity.

For example, the absolute probability that I will be hit head-on by a car at a closing speed of over 30 kph while riding my bicycle is (I hope!) low. However, the conditional probability that I will be hit head-on by a car at a closing speed of 30 kph given the situation that a drunken driver has veered to my side of the road while going at 50 kph and I am 20 m from his front fender is rather high.

I think the very best you could do, if you wish to persuade people about the rationality of always following the RA, is to engage this and other decision problems that might be presented to you, and persuade those who are sceptical (such as I am, and bsieker and joernstu are) that following the RA is the best solution in each and every one of those situations. Personally, I doubt you can do that. But it may be well worth a try, to see where it succeeds and where it fails.

PBL

[punkalouver]

Unfortunately I don't have time to reply for the next few days. I don't think there is much more that I can say than has already been posted by myself already. I suggest that anyone reading this thread, read our posts for the last few days and decide for themselves what is best and why we are trained the way we are(and what the legalities are)instead of this thread going on
forever, basically saying with us all saying the same thing in a different way.

Perhaps I have misunderstood your whole arguement. I believe I did say that there is a risk of collision in the scenario that actually happened that night but that it is an odds game which I believe you agreed with. Following the RA reduces your chance of collision especially at high altitude on a dark night where VFR traffic is a non-issue and all aircraft are or should be under ATC control.

I have mentioned a disaster(over Europe), a near world's worst disaster(over Japan) and linked to a Eurocontrol newsletter that gives multiple examples of closer than necessary near misses due to not following the RA. That is in reality all I can do to prove my point. Perhaps you could prove yours by showing me some articles where lives were saved because the pilots ignored their RA. I suppose you will point out that the DHL pilots would be alive today if they had ignored the RA. But I don't think that is enough evidence for us to realistically consider doing this as a regular procedure or consider it at all.

All I can do is ask that everyone out there FOLLOW THE RA.

Thank YOU

[PBL]

Originally Posted by punkalouver

I don't think there is much more that I can say than has already been posted by myself already

OK, noted. Thanks for your contribution.

PBL

[joernstu]

And fortunately people who behave like this(ignoring the RA) are few and far between, although there appear to be three on this thread.

I can't remember advising anyone not to follow a RA, but surely you will prove me wrong? Following an RA you get may well be the best action for you, but as TCAS cannot solve every situation, you cannot be sure, that your action will save your life (and your passengers).

I cannot decide, what kind of action would induce the lowest risk - following an RA, not following an RA or flying without TCAS at all as the basis for my risk evaluation would be episodes. I think the same goes for you as your information basis is still only episodes.

[ATC Watcher]

Punkalouver :

it is common to turn our transponder(TCAS) to TA only after an engine failure due to performance limitations.

This is new to me, and a bit surprising coming from you, as you referred to the Eurocontrol ACAS bulletins that has as first item in their do and don't list the following :

To maximise the safety benefits and operational compatibility with ATC,ten fundamental dos and don’ts must be observed:
1- TCAS II must be operated in RA mode to provide full safety benefits

That aside , I find it a pity that you do not want to engage in the debate. The point many of us have ( and I am not working for or with PBL) is that the system is still far from perfect and even if you follow the rules , it will not protect you at all times. The GOL and DHL crew if they were still alive would sadly confirm this.

The sense reversal logic currently does not work and could induce a collision while trying to prevent one.

The interaction between human-ATC and automation -TCAS is still unclear for many , especially when ATC acts before the RA, and following the RA will mean acting against the current ATC clearance.

For those 2 reasons alone, following the RA is no guarantee that a collision will be prevented, because it needs 2 willing partners and you do not know what the other is doing.

Following your argumentation , would you be for the coupling of TCAS to the auto-pilot ? That would solve many problems wouldn't it ?

[punkalouver]

Concerning turning the transponder to TA after an engine failure...

Originally posted by ATC Watcher
This is new to me, and a bit surprising coming from you

Why are you surprised. This is our SOP and the SOP of many airlines if not most. Perhaps your training department felt it was not necessary to inform you of this.

Originally posted by ATC Watcher
That aside , I find it a pity that you do not want to engage in the debate. The point many of us have ( and I am not working for or with PBL) is that the system is still far from perfect and even if you follow the rules , it will not protect you at all times. The GOL and DHL crew if they were still alive would sadly confirm this.
The sense reversal logic currently does not work and could induce a collision while trying to prevent one.
The interaction between human-ATC and automation -TCAS is still unclear for many , especially when ATC acts before the RA, and following the RA will mean acting against the current ATC clearance.
For those 2 reasons alone, following the RA is no guarantee that a collision will be prevented, because it needs 2 willing partners and you do not know what the other is doing.
Following your argumentation , would you be for the coupling of TCAS to the auto-pilot ? That would solve many problems wouldn't it ?

Hmmm, I do see that you are from Germany as well. Anyways, I don't see the point of debating any more because I really have nothing more to add to what I have already said several times(Same with the others in my opinion). The thread is is just going on and on with myself and PBL and his coworkers(or aliases) repeating the same thing in different words. We have made our points.

You are absolutely right that you are not protected at all times, TCAS is not perfect and could induce a collision and I have never denied that. However the odds are in favour of you to FOLLOW THE RA.(See how I am repeating myself again).

The Gol accident involved non-functioning equipment on one the aircraft. I don't think it proved anything in terms of the whole TCAS design except what we already know; that won't provide any protection from an aircraft with no operating transponder whether a Piper Cub or a high level jet.

No opinion on autopilot coupling as I haven't heard arguements for and against this idea.

Originally posted by joernstu
I cannot decide, what kind of action would induce the lowest risk - following an RA, not following an RA or flying without TCAS at all as the basis for my risk evaluation would be episodes

Fortunately for us, you are not flying an airliner and I suspect you would be demoted or worse if you followed through on this opinion in such a scenario.

Originally posted by joernstu
I can't remember advising anyone not to follow a RA, but surely you will prove me wrong?

bsieker(one of your like minded coworkers) in effect said so in my opinion in post #101 when he said the Tupolev crew's decision was rational which started this whole debate on the last two pages. And you strongly hint at it right after you ask me to prove you wrong by saying...

I cannot decide, what kind of action would induce the lowest risk - following an RA, not following an RA

[CDN_ATC]

Is there a website which currently lists all TCAS versions in use, their abilities (TA or RA, Climb only or turns) etc?

[ATC Watcher]

There are today only 2 TCAS in operations I and II.

TCAS I ,is a receiver that only gives TAs . description and users manual here : http://www.seaerospace.com/bfg/tcas791pg.pdf

TCAS II : the mandated system that issues TA and RAs in the vertical sense .
There are currently 2 software versions in use : 6.04A and 7.0 .
main differences > 7.0 is RVSM compliant and reduces the threat /sensitivity levels (margins ) to eliminate false alarms.
7.0 is meeting ICAO ACAS SARPS, 6.04A does not.
For a description and user manaul see here : https://www.bendixking.com/servlet/c...499-0000_7.pdf

The horizontal RA s ( i.e. turns ) was planned with a so called TCAS III but R&D on it was abandoned long ago, and it is very doubtful that they will restart.

[ATC Watcher]

Punkalouver :

cconcerning turning the transponder to TA after an engine failure you said...This is new to me, and a bit surprising coming from you

Why are you surprised. This is our SOP and the SOP of many airlines if not most. Perhaps your training department felt it was not necessary to inform you of this.

I doubt this is the SOP of many airlines . Are you going to tell us that if you have an engine failure near Max Alt in dense continental airspace , and have to perform an emergency descent you are going to degrade your TCAS to TA only during he descent ?

Hmmm, I do see that you are from Germany as well.

I am not from Germany, but that does not constitute a burden is it ?

Most contributors here are from Germany perhaps because the critical knowledge is there : Lake Constance and the German BFU are located there and the University of Bielefeld has a chair on automation that is unique ( and they have for instance overturned many stones in the Airbus garden in the past).

[bsieker]

Originally Posted by joernstu

I can't remember advising anyone not to follow a RA, but surely you will prove me wrong?

Originally Posted by punkalouver

bsieker(one of your like minded coworkers) in effect said so in my opinion in post #101 when he said the Tupolev crew's decision was rational

I would have let it rest, but since you seem to have misread my post, I will explain it again.

Describing a course of action as a rational choice does not mean that it is the only choice, nor necessarily the only rational choice, nor necessarily the best choice, nor even that the actor(s) involved acted rationally when choosing that option.

The scenario which I described in more detail, in which following ATC was the only rational choice, is the scenario in which that crew did not receive an RA at all, because of inop TCAS.

The point being that all parties following their respective rational (in this case, even best) choices, one of which is following the RA, the other, in the absence of an RA, following ATC, (courses of action, I take it, you would approve) and yet they end up in a collision.

The problem is not so much following an RA, but the poorly-understood and unregulated (non-)interaction between TCAS and ATC.

This, as we have seen, may lead to collisions, although and because one crew follows its RA.

You say that you always follow the RA because the risk is lower than following ATC, where both are in conflict, but without giving even the hint of an argument for it, in the face of strong arguments (i. e. Ueberlingen) to the contrary. The Boeing pilot could not know if the Tupolev's TCAS was inop, or if they were just ignoring its RA.

If you misconstrue this as me generally recommending not following RAs, I cannot help it.

I am merely pointing to what I think is a big problem with collision avoidance in the larger system "air traffic", not looking at TCAS as an isolated system, but at the larger socio-technical system, comprising (at least) of two or more TCAS systems, two or more flight crews, several air traffic controllers, primary and secondary radar systems, ATC software, ...


Bernd