Human factors and Glass cockpit errors

[Skaz]

Hi all,
I am writing a college paper on the errors happening on the flightdeck of the big jets (and little jets) with glass cockpits. Specifically the human factors induced errors generated by the introduction of glass cockpits to crews unfamiliar with their operation.

While there are many accident websites to dig into, finding specific information is rather difficult.

Any info re academic or actual personal experiences would be greatly appreciated.

Skaz

[BlueEagle]

I worked for an airline in the UK that changed equipment from the classic instrument presentation to glass, this meant that both Capt and F/O were new to the technology.

What we soon realised was that there was a tendency to spend too much time head down trying to figure out the FMC (Flight Management Computer) and not enough time spent flying the aircraft. Obviously this was spotted early on and a positive discipline was applied to prevent it from happening. There is a strong temptation when one's colleague says, "What is it doing now?" to want to look down and help out, particularly in the early learning stages. Fortunately the learning process is quite fast once pilots are on line and the SOP of only the non-handling pilot making changes to the FMC whilst the handling pilots flys the aircraft helps to ensure a safe operation.

There is a tendency amongst some pilots to spend far too much time head-down in the FMC during critical and busy stages of flight, such as departure and arrival. When being radar vectored to an ILS intercept or a visual final it is necessary for both pilots to concentrate on the flying, R/T, configuration changes etc. and leave updating the FMC alone, the same with departures.
Whilst under positive radar control for heading, height, speed etc. the information from the FMC is not always relevant and using back up information such as DME and the ADF needles can be much more helpful.
Finding the right compromise between the automation offered by the Glass cockpit and alternative sources of information available to the pilots is important, I believe, in helping to ensure the maximum attention of the pilots in the appropriate direction at any one time.

These are my personal views, by the way, and offered as such! :)

[Skaz]

thank you sir, this kinda thing is exactly what I wanted personal stories from pilots who fly it.

[safetypee]

Try the following:-

European Collaboration On Transition Training Research for Increased Safety (ECOTTRIS)
http://www.nlr.nl/public/hosted-sites/ecottris/index.htm


BASI Advanced Technology Aircraft Safety Survey Report
ISBN 0 642 27456 8 June 1998
http://www.basi.gov.au/ but this site may have been superseded by CASA


FAA The Interfaces Between Flightcrews and Modern Flight Deck Systems
June 18, 1996
http://www2.faa.gov/avr/afs/interfac.pdf


Lufthansa (Capt Dieter Uchtdorf and Capt Peter Helt) surveyed the B737-300 and A310, the results were published in a special edition of their Flightcrew Info magazine 1989

[Groundbased]

you can also try this:

http://www.rvs.uni-bielefeld.de/publications/Incidents/

This site contains accident reports you can download and many good references for the
human/technical interface in the cockpit.

Also if you do a google search for the columbian authorities report on the AA 757 crash at Cali
you will find it explores these factors in relation to this accident in depth.
There are also good references to books about decision making and risk strategies in this
report.

There was recently a thread on this forum discussing the use of speed tapes in the PFD and how they are interpreted which provided many interesting views on this subject.

[Skaz]

wow...thanks a lot for all the gen, it is really appreciated.

:D

[Kotare]

You might also like to try the following: http://www.bluecoat.org/
There is an extensive library of URL's etc on the subject.
cheers

[Sick Squid]

Interesting question. I've spent my entire professional aviation career in glass machines of one shape or other, so have on one level no comparisons with conventional aircraft. However i have noticed some fundamental differences in the glass philosophy as presented to the pilot, and also some common traits come through.

First thing.. there is a world of difference between the PFD/ND presentation of the 744/777/320 and the "conventional" EADI/EHSI-type that you find on the 757/767/737. The scan on the latter is very similar to an analog aircraft's conventional 'T' pattern, with the exception that some of the more useful glass add-ons such as wind vectors and digital inertial-groundspeed are gathered together on one of the "glass" displays, increasing the concentration of information in those areas. However, i think it's fair to say that in terms of primary flight instruments, the mix of glass and conventional is relatively easy to come to terms with, leaving aside the FMC and engine instrumentation considerations.

The PFD/ND presentation on the other hand, is where the nature of the glass cockpit issue starts to surface. The information derived from the 'T' scan is now very compacted, and on some types enclosed within the entire PFD; 777 and 320 with the partial compass rose at the bottom being examples of that. Also the ND with the ability to display map with superimposed raw data needles provides a valuable resource, but also one that can very easily become "cluttered" requiring careful choice of what is "required" on the day... I am a firm believer in the need to at all times have a raw-data source on the nav display; this is in direct variance with the philosophy on one type I flew where it was recommended that if the GPS was fully updating, then there was no need to display needles unless required for the approach/procedure being flown. Always gives the route-checkers something to debrief on, they love that..... took the point on board, but still had the needles up, my choice.

Paradoxically, one of the greatest benefits of the glass cockpit is one that was actually designed in as an elastoplast-type fix for a known perception problem. I am referring here to the speed-trend vector on the digital tape ASI section of a PFD.... sometimes also appended to an EADI as an adjunct to the conventional ASI next door. The vector was initially provided because humans just dont cope well with lines of numbers moving, and there is a potential illusory effect where the direction of travel of the tape can appear to reverse, i believe.

The vector provides the indication of where you will be in ten-seconds (generally) given current energy state. Invaluable in certain cases.. however it simply must have a form of damping added to it, otherwise it over-reads quite dramatically. Humans being as we are, we can compensate for that, but having flown both the over-sensitive vector of the 737, and the well damped 744 (and to an extent 777) as well I am in no doubt which is the more beneficial. Indeed, I have heard many on "conventional" glass types talk of "ignoring" the speed-tape and the vector, simply because it is not damped enough.

When it comes to the other area of glass machines, the engine indication section another whole series of issues appear. Firstly, certain types provide schematic indications of systems; excellent to a level, however unless there are actually flow-integrators fitted to provide an accurate indication of fluid travel on the relevant displays, then they can be rendered next to useless. Types that do that are 320 and 777, whereas the 744 derives flow indication from switch position. You have to question to a certain extent the relevance of such.. however, as long as it is hammered home on the conversion course, and a modicum of intelligence applied then shouldn't really be too great a problem.

Also where an electronic checklist is provided, it simply MUST be relevant, accurate, and comprehensive enough to do the job. I witnessed the development of this on the 777 from the early incarnation where a non-normal checklist requiring secondary actions (eg.. engine failure causing electrical generator failure as secondary) required one to individually dig out the secondary checklist from the system at the point it appeared, cancel it, then return to the primary checklist. A software revision fixed that, and the difference in workload and therefore capacity was eminently noticeable.

This manifests itself (in my opinion) on the 320-series where the discipline required to correctly run the non-normal checklist is required over and above the level that should be necessary... it can be argued that it in fact detracts from resolution of problems by putting another notch on the workload at a time where it is already increased. Also, certain fundamental procedures there are potential traps... witness the fuel-leak Azores problem, and also the smoke procedures, both of which actually require a paper QRH to complete safely.

Bit of a long post, but it is a very interesting subject, and one I've certainly had to take time out from and think through before to ensure i do the best I can. Hope it helps.

£6