Analysing FlightData

[paull]

Does anyone (Airbus, Boeing, large carriers) systematically or on an ad-hoc basis gather and analyse flight data other then when there is a crash?
in the recent Airbus overrun there is much discussion about just how unusual it is to forget to pull the throttle back on landing. Well how often does it happen? It seems to me if we want to systematically drive improvement then you need to know the answers to these questions without waiting for people to die (a few times).
Some 25years ago the company I worked for would pull error log data from computers that gave folks like me driving reliability inmprovements a large database that we could trawl through to see how often a given event happened. If something was not in our usual snapshot of data we could send an ad-hoc request for different specific one-off data to be sent to us.
So, how often does someone forget to put a throttle back to idle, and when they do, how long does it usually take them to work it out?
Obviously there are a ton of other benefits also, such as checking if people have upgraded their software to the latest rev.
If this is being done properly then for any fault (in the larger "man-machine sense") Airbus or Boeing would be able to immediately quantify the impact and potential frequency and use this as a process for driving continous improvement. Please tell me it is already in place.

[Bullethead]

The mob I work for monitors our quick access recorders for certain selected parameter exceedences and occasionally procedures are modified if a negative trend is developing.
I don't think nit picking stuff like not having the throttles completely closed at touchdown would be monitored. Personally I've done two landings in the last year where I had roughly climb thrust set at touchdown due to the sudden and unexpected disappearance of a 30kt headwind component in the flare. Both landings went on smoothly but certainly wouldn't have without all that thrust.
Any exceedance is de-identified so indviduals can't be nailed but if a trend is developing then things get changed.
Regards,
BH.

[Mad (Flt) Scientist]

FOQA (the FAA term, Flight Operational Quality Assurance) does what you describe:

Flight Operational Quality Assurance (FOQA)
FOQA collects and analyzes digital flight data generated during normal line operations. FOQA programs provide greater insight into the total flight operations environment. FOQA data is unique because it can provide objective information that is not available through other methods. The information and insights provided by FOQA can improve safety by significantly enhancing training effectiveness, operational procedures, maintenance and engineering procedures, and air traffic control procedures.

Others refer to such programmes as "Flight Data Monitoring".

Right now I believe these programmes are technically voluntary, though if memory serves there is some point where they'll be mandatory. And any carrier operating under a "Safety Management System" (SMS) may find a FOQA/FDM programme a mandatory part of such a system.

Bear in mind, though, that identify a problem is one thing; devising a solution is something else, and forcing its implementation something else again. Note that OEMs have no legal means to impose design changes on operators; only national airworthiness authorities can do that.

[paull]

Thanks Mad(F)S
FOQA (the FAA term, Flight Operational Quality Assurance) does what you describe:
As you say finding a fix is something else but it grieves me think that for every one of these that happens on a wet short runway we have probably seen 7 on a long dry runway that never came to anything.
I know that every problem we ever saw operationally had already been seen (in our case) in manfacturing at some earlier stage but someone thought it was a one-off. Once they started saying "We have seen this, is there any evidence that it is more than a one-off?" then we really started to work the right improvements. Actually we shut it down after 2 years because we had got all the low hanging fruit and there were more profitable areas to focus on, of course in our case we were just trying to save labour and spare parts not lives.

[paull]

BH
I don't think nit picking stuff like not having the throttles completely closed at touchdown would be monitored
I explained that badly, I was thinking of this current case where one throttle was in reverse and the other still at CLB, I take your point about power on landings.
Good to see that the data stays anon., I had the workers council in Germany threaten me with arrest when they misunderstood my data gathering objectives during one study! My fault, poor communication.

[XPMorten]

http://www.flightdataservices.com/

M

[PJ2]

paull;


If this is being done properly then for any fault (in the larger "man-machine sense") Airbus or Boeing would be able to immediately quantify the impact and potential frequency and use this as a process for driving continous improvement. Please tell me it is already in place.


In a word, yes.

As you have seen by the responses already posted, there are organized "audit" programs in place called FOQA (Flight Operations Quality Assurance), or FDA (Flight Data Analysis).

With some programs, the de-identified data (no date, no flight #) is kept instead of destroyed, so that questions can be asked of the massive (in the terabyte range) database. This has been mentioned in the thread already. The technique is to construct a question (using PERL or other query language) to "trawl" (good word) through the data. Ordinarily it would take weeks with such a huge database but there are ways of doing this quickly.

in the recent Airbus overrun there is much discussion about just how unusual it is to forget to pull the throttle back on landing. Well how often does it happen? It seems to me if we want to systematically drive improvement then you need to know the answers to these questions without waiting for people to die (a few times).


We are already now asking this question of our stored data.

[discountinvestigator]

There is an implied requirement for FOQA/FDR for many countries across the Western world. Each carrier owns their own data, not the manufacturer. It may be possible to allow one carrier's data to be used in a cross-carrier analysis, as implied by PJ2.

Failure to have a FOQA programme may consitute unreasonable behaviour as it is "reasonably practicable" to fit this equipment to modern jet transport aircraft.

However, I see many bad cases of FDR/QAR analyses by accident investigators out there. I see many cases where the FDR data has not been used to identify the real issues.

Whilst data are very useful, one should never forget the basics in aviation. We could save so many more lives if the airports, ATC, AIS and met service providers and their regulators did their jobs properly as well.

Still, as carrying an FDR for analysis on a flight as hand baggage is regarded as a major security risk by some governments as it is too big to fit in the standard bag gauge, I am sure that we will lose a few more in transit. I have only lost one so far with the current restrictions... :ugh: Perhaps the Heathrow baggage system might manage to crush a few more in the name of security?

[Mad (Flt) Scientist]

One problem with FOQA as currently set up is, as you note, that the OEM is very much out-of-the-loop. This can lead to some very strange conclusions by those reviewing the data for SYSTEMS problems, as often they have not much more to go on than the FCOM, which is not a detailed design decoument by any stretch of the imagination. (I've seen at least one formal presentation by an airline FOQA team of the "causes" of an incident which was clearly based on a very rudimentary understanding of the system design)

(By the same token, of course, the OEM is very much isolated from actual operational practices, so we can be pretty much at sea when trying to understand crew behaviour.)

The reality is that only a truly multidisciplinary "accident investigation" team can hope to understand a serious incident or accident in most cases. As a tool for monitoring adherence to SOPs and routine issues FOQA is a great idea. And as a tool to improve system reliability predictions and catch systems issues earlier it'd be great if the OEMs got something too. But neither operators nor OEMs should be in the accident investigation business, except to assist the REAL investigators.

[paull]

Some interesting comments on how best to do this and how to get it widely installed. Perhaps a few conclusions based on my own experience, albiet with a self policing OEM!
We do not have all the data
My principle was not to ask anyone for any more data unless I had done all I possibly could with the data I had. It is easy to ask for more and more and then just let it sit there. You need to be sure that all the stakeholders are seeing the results and getting some benefit. When you say to them "I have done this, but if only I had... " then you will get more data.
Not everyone does good analysis
I think you still have to throw it open to all. Over time the DataAnalyst gladiators will develop some mutual respect and the cowboys will be shown to be such.
OEM's are not the ones to do this, they are not independent enough
I worked for an OEM and was never under pressure to do anything other than the right thing. I believe as professionals we would never bend the data. The only concession I made to my employer was to give a "heads-up" warning of bad news before it was pubished, but then I would do that for any participant in a program.
How do we get such a program mandated?
1-Get Airbus & Boeing to agree that they at least are fully committed.
2-Then they agree that all aircraft are supplied with whatever equipment is needed be it on-board or ground based at no extra cost.
3- OEMs give a discount on maintenance charges or parts or whatever for airlines who agree to supply the data on a regular basis. (We discounted maintenance charges to any customer that agreed to connect to the RemoteDiagnosis Centre)
4- The analysts in the meantime work hard and make sure that what they produce does show tangible benefit to the participants.
On this last point, I know there is much discussion implying savings on insurance premiums for those who demonstrate a good Safety Management Program but you guys have a chance to plenty in this area.
- Although hull-losses are thankfully low I am sure you could develop a model that showed how many times a given OEM or airline is in the position of having all but 1 ,2, or 3 holes lined up in the swiss cheese. Actually this gives me a few ideas...........
5- OEMs and Airlines arrange job-swaps between managers of appropriate groups. This for us was the big factor in the end. I never got to ground a fleet but I did get to the point where we looked to be about to put a worldwide manufacturing hold on shipping our latest high-profile product. At times like that people on both sides need to know they can trust each other and that only comes after a certain degree of common suffering and no small number of beers!