Go Back  PPRuNe Forums > Flight Deck Forums > Rumours & News
Reload this Page >

TAM A320 crash at Congonhas, Brazil

Rumours & News Reporting Points that may affect our jobs or lives as professional pilots. Also, items that may be of interest to professional pilots.

TAM A320 crash at Congonhas, Brazil

Old 6th Oct 2007, 19:36
  #2701 (permalink)  
PBL
 
Join Date: Sep 2000
Location: Bielefeld, Germany
Posts: 955
Originally Posted by John Marsh
Perhaps this only seems 'primary' because all the other holes in the Swiss cheese lined up.
The Swiss cheese is a wonderful metaphor, otherwise it would not have persisted for thirty years.

But I wonder sometimes whether the people who use the metaphor actually remember the theory behind it. The cheese has layers. What are the layers called? And how does the theory apply to the Congonhas accident, in detail?

PBL
PBL is offline  
Old 6th Oct 2007, 19:40
  #2702 (permalink)  
 
Join Date: Nov 2001
Location: Somewhere in the Tropics UTC+7 to 9
Posts: 450
The primary cause is the primary cause which made the occurence an EVENT. Whether the event is an accident or an incident, is a different matter.

Runway conditions were contributory into making the event into an accident. It was not the primary cause on why the occurence became an event. It could be biggest contributory factor that made the event turn out as an accident.

What made the landing became an event was the TL positions.

So let's all forget to retard the TLs and blame the runway? Kidding...

PBL, 1.70.90 was in interesting read! Thanks mate!

PK-KAR
PK-KAR is offline  
Old 6th Oct 2007, 19:43
  #2703 (permalink)  
 
Join Date: Jul 2002
Location: Confusio Helvetica
Posts: 305
310: I'd strongly suggest you reconsider your notion of aetiology.

In the example you give, your argument is:
Tens of thousands aircraft land at the CGH, very few finished off the runway.
This one finished off the runway.
A major distinguishing factor is the TLA position.

Therefore, the TLA position is the "primary cause".

That's weak. PBL's counterargument is:

Numerous A320s have attempted landings with the TLA in the wrong position
This one is one of the few that finished off the runway
A major distinguishing factor is the shortness of the field

Therefore, CGH is the "primary cause".

---
In history, we would call such an oversimplification an instance of the single-cause fallacy.

But the problem is, accident investigations are not pure historical endeavors. The investigators are charged with not only determine how an accident came about, but in making recommendations so it does not happen again. Their findings, and their recommendations carry considerable economic, legal and political weight. So, in arguing their point, they put forward a scheme of "primary" and "contributing" causal factors that don't seek purely to explain the event, but rather intend also to provide the foundation for their recommendations. "Primary/contributing factors" are the basis for arguments, not thorough explanations of the event. In other words, what's happening is that some of the causes - actions, inactions, conditions, circumstances - are being emphasized for purposes that are beyond simply explaining "what happened".

We had some bickering over the definition of the term "Accident", including folks citing Websters and lucky part 13 of the ICAO. But whatever the case is, I hope we can agree that most aviation accidents are the results of several factors: the "holes in the Swiss cheese", as y'all like to call them. A chance or fortuitous configuration of not necessarily related circumstances result in an accident. Well, every hole in the cheese is a causal factor.

Both the TLA remaining where it was and the short, wet runway are necessary conditions here. That makes them causes. Neither of them alone is sufficient to cause an accident.

Now, once you identify the causes, then you can go to the business of making recommendations, and cheerfully selecting whichever ones support your political, economic or safety agenda, and calling them "primary" and "contributing" causes. But, in themselves, and in the event, "primary" and "contributing" causes have no meaning. There are just causes, and sometimes effects.
DingerX is offline  
Old 6th Oct 2007, 19:49
  #2704 (permalink)  
 
Join Date: Feb 2005
Location: flyover country USA
Age: 78
Posts: 4,579
I like the "holes in the Swiss cheese" analogy. If any one of the contributing factors were absent, these folks would still be alive.
barit1 is offline  
Old 6th Oct 2007, 21:54
  #2705 (permalink)  
 
Join Date: Mar 2002
Location: Florida
Posts: 5,382
Bari1

I like the "holes in the Swiss cheese" analogy. If any one of the contributing factors were absent, these folks would still be alive.
I'm not sure I can agree with this. But just in case I'm missing something could you put forth a list of contributing factors that you believe are in a direct line, rather than trunk lines?

The failure to retard was in my mind an egregious factor which may have overridden some other suspected contributing factors bandied about in the overall discussion.. But I will await seeing your list.
lomapaseo is offline  
Old 6th Oct 2007, 22:04
  #2706 (permalink)  
PBL
 
Join Date: Sep 2000
Location: Bielefeld, Germany
Posts: 955
Originally Posted by barit1
I like the "holes in the Swiss cheese" analogy. If any one of the contributing factors were absent, these folks would still be alive.
That is indeed so. The weaknesses with the analogy are that
1. There may be more factors which, if they were absent, those folks would still be alive; how do we know we have them all?; and
2. The factors which fulfill your criterion (namely those things which, if they had been absent, those folks would still be alive) do not necessarily respect the property of one-factor-to-one-slice of the cheese model. There may be many factors in one slice, and no factors to another.

P.S. My favorite Swiss cheeses are (good) Gruyere, and Tomme, and I don't like Emmenthaler much (that's the one with the holes).

PBL
PBL is offline  
Old 6th Oct 2007, 22:42
  #2707 (permalink)  
PBL
 
Join Date: Sep 2000
Location: Bielefeld, Germany
Posts: 955
Just to put a few thoughts more about causes into circulation:

Suppose I have a job which requires me to walk 20 meters along a horizontal bridge made out of a tree trunk of 20 centimeters diameter.

The SOPs say how I am to do it.

Sometimes my pole lies on firm, sandy ground.

Sometimes my pole is suspended 1 m above a marsh.

Sometimes my pole is suspended 100 m above a boiling geyser.

General human factors stuff tells us that one in a thousand attempts I will fall off. Of course, there are some pole walkers that never fall off. And there are some pole walkers that falls off one in a hundred attempts. And 80% of pole walkers believe they fall into the first category, not the second.

Somebody walks on the pole 100 m above the boiling geyser, falls off.

Some of his fellows say: he's obviously not "first category" material. Shouldn't have fallen off. We do our best never to fall off and cannot understand why he did. We advise all our fellow pole-walkers not to fall off. Don't do it. Ever.

Some of his fellows say: well, he certainly shouldn't have fallen off, but he did. We don't intend to, and do our best not to, but maybe his shoes weren't quite right; they are not as good as we think ours are. And maybe when he wobbled a little and tried to correct his stance, his shoes deformed and he lost his balance. We won't use those shoes.

Some of his fellows say: you fall off a pole generally once in a thousand attempts. Our career is comprised of 500 pole-walkings. We have a one-in-two chance of falling off. So we'll go on the firm, sandy ground. We'll even go on the 1m above the marsh. But there is no way we are ever going to do the 100m above the boiling geyser. Forget it.

Question: who has the surest way of avoiding killing himherself?

Another question: Is the "primary cause" the failure to keep balance; or is it the wrong shoes; or is it the situation (100 m above the boiling geyser)? Or is it all three? Or is it some combination?

A further question: who is "responsible" for the accident?
The pole walker? Hisher shoe manufactuer? Or his employer who requires himher to walk 100 m above a boiling geyser?

We have seen examples of all three of these causal points of view in this discussion. Who is right? Is anybody right? Or are they all missing understanding something about causes?

PBL
PBL is offline  
Old 7th Oct 2007, 00:20
  #2708 (permalink)  
 
Join Date: Sep 2007
Location: Timbuctu
Posts: 27
As a consequence of PBL's theories:

A murderer should not be punished... Without the life he took from the guy he killed there wouldn't be that specific murder...

A drunk driver that hits somebody's else car by driving in a highway 100mph above maximum speed and kills everybody in the vehicle is not guilty.. After all why should one expect that the victim's car being driven within the rules should not be listed as a cause too ?

Jesus Christ, many thousands, maybe millions of flights have landed safely on that runway, including with A320s, and yet a TL reading above idle after touchdown is not more relevant than the runway length ?

It is an elementary thing that even children do - correlate things and weigh the importance of an event in the materialization of another one -

This is how science is done by the way. It is the essence of the scientific method.

Wake up, please...
aviadornovato is offline  
Old 7th Oct 2007, 01:10
  #2709 (permalink)  
 
Join Date: Jul 2002
Location: Confusio Helvetica
Posts: 305
To take your second example: most drunk drivers get home safely. In fact, I'm guessing that something on the order of only one in ten thoiusand drunk driving sectors ends with anything but an arrival at home.

What some folks aren't seeing is that, while "cause" is a neutral assessment of an event - even Aristotle, who gave us the first formal theory of causation, would argue that the victim was a cause of the murder - "responsibility" implies a judgment.

Before you judge, you have to evaluate the whole causal system. Then, when you run into a stupid airmanship mistake, you need to look at the circumstances that caused it. Unofficial procedure trying to squeeze out those phantom fifty-five meters? People pouring gasoline all over the place? Unsafe conditions breed unsafe behavior. But those are my judgments.
DingerX is offline  
Old 7th Oct 2007, 01:11
  #2710 (permalink)  
 
Join Date: Sep 2007
Location: Portugal
Posts: 39
My last post was a bit rushed, now I have more time. I have to stress that Iím not discussing this *primary cause* issue just for the sake of argument, I think the categorization of these factors is important in the understanding of accidents, and thus for flight safety.

DingerX wrote:
Numerous A320s have attempted landings with the TLA in the wrong position
This one is one of the few that finished off the runway
A major distinguishing factor is the shortness of the field

Therefore, CGH is the "primary cause".


I realy cannot agree with this way of putting things. The simple fact that one can make a association between facts doesnít make it valid *per se*. Landing on a short runway isnít a wrong action, unless you are out of the parameters to accomplish it (as Lemurian said, it is possible that the a/c was at a weight that required more LDA than the available, and that, in fact is wrong). Failing to retard the throttle is a wrong action *in any runway*. So, in this light, maybe you can tell me wich of the facts is more relevant.

PBL wrote:
The thing is, 3Ten, if I may phrase it so, that you haven't thought about these things very hard yet. You're applying the Counterfactual Test selectively, but you haven't said what your selection criteria are, or justified these criteria.

Let me propose three simple resolutions to this discussion. Either
1. I promise not to tell you how to fly your large commercial aircraft if you promise not to tell me how to think about causality; or
2. I give you a rather long reading list in the logic of causality and its practical application in analysing accidents, and you read it all; or
3. You take our course in practical analysis of accidents using WBA.


Starting by the end, for number 3, maybe one day Iíll do it, itís na area that really interests me, unfortunately not the time yet.

Maybe this discussion is being affected by a difference in point of view. Iím a pilot, and consequentely, my priorities in analising a pilot error related accident are as follows:

Human failure related with basic airmanship principles: Our profession requires that we mantain a great ammount of knowledge of various cathegories. The basic airmanship is as the name states, the base that we have to achieve the safe conduct of flight, and exactly the type of knowledge that we have in the most *ready to use* condition, is (or shoud be) already built in our mind, and isnít all that vast. So, it is our best chance to break the accident chain. I understand that these factors, beeing very meaningfull for the pilots, are less meaningfull for other parties involved, because these other parties donít have as much to learn from this erros.

Ohter causes relate do various levels of our knowlege, levels that are in different states of availability during flight operations. These causes are much more vast in their scope, and cannot be all recorded in *ready to use* condition by us, so we have to use them to build a general awareness state, to develop a more critical attidude during the decision process. However, this causes may be much more relevant to other parties, as they relate to these partieís areas of action, say ATC procedures, airport and aircraft design, maintenance procedures, etc. Those parties may learn a lot more from these causes than from the former ones, but for the pilot, they may be less relevant.

Other factors, although relevant in their role in the accident, may be cathegorized by the pilot as *nice to know*, as they fall in areas totally out of the pilotís scope and influence, but still be very important to other parties.

We all know that accident investigation doesnít exist just because of pilots. But, when pilot error accidents are investigated, typically basic airmanship failures are appointed as primary cause, they donít fall on other cathegories. Iím not na accident analist, and Iím sure PBL can correct many of my choice of words in this post, but I hope I could convey what I think about this. When Iím airborne, I cannot change an aircraft design, or a runway condition, but I can do my best to overcome this problems and try to avoid an accident. So, we have different priorities.
3Ten is offline  
Old 7th Oct 2007, 03:31
  #2711 (permalink)  
RWA
 
Join Date: Sep 2007
Location: Melbourne
Posts: 180
Originally Posted by PBL
But if they had had a longer runway, with escape areas, would we be here discussing this?
Originally Posted by DingerX
Numerous A320s have attempted landings with the TLA in the wrong position
This one is one of the few that finished off the runway
A major distinguishing factor is the shortness of the field
Sorry, PBL, DingerX, but I think the logic of those statements is just plain awry.

Congonhas is the fourth recorded serious occurrence of this type - and the major points of similarity are:-

1. Aircraft was A320.
2. One reverser was inop.
3. One TL was wrongly-positioned.
4. ALL aircraft eventually swerved off the runway BEFORE the end.

There is no apparent correlation as to runway length. At Bacolod the length was 5,932' (i.e. shorter than at Congonhas); at Phoenix 11,500'; and at Taipei 9,200'. Further, the availability or otherwise of run-off areas would appear to be immaterial since none of the aircraft reached such areas before going out of control and veering off the runway altogether.

In the case of Congonhas, in my opinion:-

1. The 'condition' of the runway (newly-surfaced, ungrooved, and wet), and its limited length, are highly likely to have contributed to the SEVERITY of the accident.

2. However, given that large numbers of aeroplanes (including other A320s) had landed at Congonhas without incident in the same conditions on the same day, and taking into account the evidence of previous similar (heck, almost IDENTICAL) occurrences on much longer runways, neither its condition nor its limited overall length would appear to be in the same league as the wrong TL position, the inop. reverser, and the non-availability of both ground spoilers and autobrakes as 'causes'.

Last edited by RWA; 7th Oct 2007 at 07:14.
RWA is offline  
Old 7th Oct 2007, 04:23
  #2712 (permalink)  
 
Join Date: Feb 2005
Location: flyover country USA
Age: 78
Posts: 4,579
lomapaseo asks...
But I will await seeing your list.
I don't have an exact list for this accident, nor (I suspect) does anyone else other than the parties to the official investigation.

And (as I used to tell my students) you cannot PROVE your case by analogy, but you can certainly use it as a powerful teaching tool.

The point to be made is that it's unlikely a single causative factor is the culprit.
barit1 is offline  
Old 7th Oct 2007, 05:02
  #2713 (permalink)  
RWA
 
Join Date: Sep 2007
Location: Melbourne
Posts: 180
Originally Posted by barit1
The point to be made is that it's unlikely a single causative factor is the culprit.
But it's possible (and perfectly logical) to prioritise them, barit1.

You have a glass coffee-table in your dimly-lit living-room. Hurrying to answer the phone, you forget that the table is there, trip over it, and fall. Because it's made of glass, it breaks, and you end up with severe cuts.

What would be the basic cause of the accident in that case? Would the phone call, the poor lighting, the fact that the coffee-table was made of glass, or even the fact that the coffee-table was there at all, be 'equal,' in terms of causality, with your own carelessness?
RWA is offline  
Old 7th Oct 2007, 08:28
  #2714 (permalink)  
PBL
 
Join Date: Sep 2000
Location: Bielefeld, Germany
Posts: 955
Originally Posted by aviadornovato
As a consequence of PBL's theories:
Astute readers will have noticed that I presented no "theories". I presented a situation (or a "thought experiment" as they are sometimes called) and asked questions.

The reason is that I wanted people interested in ascribing causes not only to think about the issues and give answers, but to provide some justification for the answers. Most people find it easier to do when presented with a "thought experiment" with subject matter somewhat different from that which they are used to.

The theories I use, when I use them, are largely attributed to David Lewis and David Hume. Hume, widely regarded as being the most significant thinker about causality in general in the last 250 years, proposed two criteria for being a cause. One is the "constant conjuction" criterion well-known to many. The other is the Counterfactual Test. David Lewis put logic and rigor behind the Counterfactual Test 33 years ago, and his theories have withstood significant, sustained criticism over that time. The weaknesses are well known. This doesn't satisfy all logicians and philosophers, who are looked for an exceptionless explanation of causality, but it is good enough for practical use, since if one is aware of the situations for which exceptions occur, one can identify them and handle them separately in any practical application, and still get 99% of the application handled via the Counterfactual Test.

The Counterfactual Test is also the basis of the U.S. Air Force incident analysis requirements (although not named as such) and is the semantic basis for Andrew Hopkins's Accimaps, which have been used by the ATSB to analyse the Lockhart River accident as the successor (or one of the potential successors) to Jim Reason's theories (the more precise form of the "Swiss cheese").

The Lockhart River scheme uses a model with five levels of factors:
* Occurrence Events
* Individual Actions
* Local Conditions
* Risk Controls
* Organisational Influences

There are accimaps showing the causal relations of various of these factors (whose semantics are supposed to be the Counterfactual Test) on pp201 (p224 of the PDF), 210 (233), 220 (243) and 227 (250). The most complete accimap is that on p227.

The major differences between accimaps and WB-Graphs are that:
* the accimaps summarise or aggregate the individual factors before representing them, and are intended as a presentation device; that is, they have to appear on one page and be easily readable, so the aggregation is quite significant. In contrast, WBA does not aggregate, but exhibits all the detail in the WB-Graph. We regularly construct WBGs with 80-100 nodes;
* the accimaps group the factors into the five layers enumerated above.

For those not yet aware, there is a WBG of the Congonhas accident which incorporates the factors known publically so far (and is thus necessarily incomplete) at
http://www.rvs.uni-bielefeld.de/publ...s/TAM3054.html

I will try to heed aviadornovato's advice to wake up, but it is a Sunday morning......

PBL
PBL is offline  
Old 7th Oct 2007, 08:58
  #2715 (permalink)  
PBL
 
Join Date: Sep 2000
Location: Bielefeld, Germany
Posts: 955
A thoughtful post from 3Ten.

3Ten, allow me to summarise what I think are your criteria for prioritisation. In terms of the Lockhart River categories, you are prioritising Occurrence Events (OE) and Individual Actions (IA) over Local Conditions (LC) and Organisational Influences (OI). I am not sure where you would put Risk Controls (RC).

The reasons you give is that, as a pilot, you have most immediate control over IA and then OE, and less over LC and OI.

I suspect alf5071h would argue that that leaves RC, and pilots have considerable control over RC. For example, the FSF slides presentation "Managing the Threats..." which was referenced early in this thread, spends a lot of time on how pilots can manage the risks of landing.

At Congonhas, for example, the pilots had information about the runway condition, besides what they knew themselves about the runway; they knew they had one INOP reverser; and they planned to land shorter than usual. But they were also working in conditions in which there did not appear to be SOPs that were adhered to rigorously on how to handle thrust on landing with one reverser INOP.

And, maybe needless to say, various different groups have various different interests and thereby select various different causal factors for emphasis. I prefer to have them all out there, and to make it clear that the emphasis on one set of factors rather than others is a choice independent of determining the causality. It can be made, as DingerX suggests, because there are more chances for prophylaxis; or it can be made, as you suggest, because you are an agent and those are the factors under which you have the most control; or it can be made, as others have noted, in order to attempt to avoid deleterious consequences for one's organisation (financial or political liabilities, for example).

PBL
PBL is offline  
Old 7th Oct 2007, 08:59
  #2716 (permalink)  
Per Ardua ad Astraeus
 
Join Date: Mar 2000
Location: UK
Posts: 18,581
PBL - which of your 3 'pole dancing' events was/were an accident?
BOAC is offline  
Old 7th Oct 2007, 09:04
  #2717 (permalink)  
PBL
 
Join Date: Sep 2000
Location: Bielefeld, Germany
Posts: 955
Originally Posted by BOAC
PBL - which of your 3 'pole dancing' events was/were an accident?
I only enumerated one event, namely

Originally Posted by PBL
Somebody walks on the pole 100 m above the boiling geyser, falls off.
That's the event about which I was asking the questions.

PBL
PBL is offline  
Old 7th Oct 2007, 09:11
  #2718 (permalink)  
Per Ardua ad Astraeus
 
Join Date: Mar 2000
Location: UK
Posts: 18,581
So had the chap (with dodgy shoes) fallen off the other 2..........first?
BOAC is offline  
Old 7th Oct 2007, 09:30
  #2719 (permalink)  
PBL
 
Join Date: Sep 2000
Location: Bielefeld, Germany
Posts: 955
Originally Posted by BOAC
So had the chap (with dodgy shoes) fallen off the other 2..........first?
I am not sure that I wish to expand on the example, because it detracts from the simplicity of the situation.

I think you can probably attempt to answer my questions, should you be so inclined, with the situation as I gave it. If your views depend on his having fallen off the other two first, you could say:
"Suppose he had already fallen off the other two first. Then ......."

PBL
PBL is offline  
Old 7th Oct 2007, 10:37
  #2720 (permalink)  
 
Join Date: Jul 2007
Location: the City by the Bay
Posts: 530
BOAC may have something here. Think Bacolod, TAipei, Congonhas.
armchairpilot94116 is offline  

Thread Tools
Search this Thread

Contact Us - Archive - Advertising - Cookie Policy - Privacy Statement - Terms of Service - Do Not Sell My Personal Information -

Copyright © 2018 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.