Human
Error — The Last Frontier in Aviation Safety
Powered flight has enjoyed a century of progress,
spurred on by two World Wars and those sudden leaps in technology that often
look so obvious in retrospect.
Fifty years ago military aircraft struggled to exceed Mach 1
as they juddered downwards in steep dives from the tropopause. Now we have
thousands of airliners criss-crossing the skies at cruising speeds exceeding
the best efforts of military fighters of the 1950s. These airliners have
reliable engines producing thrusts of an order greater than was available
then. The automated systems assisting the crews of these huge aircraft would
have been totally incomprehensible to scientists fifty years ago. Passengers
and crews now sit in air-conditioned and pressurised comfort without the need
for pressure suits and oxygen masks to survive the hostile environment
surrounding their semi-automatic flight.
The safety of operation of both military and civilian aircraft has
made considerable progress over the last 50 years. Going back even further in
time, the Royal Air Force flew from dawn to dusk on 8 August 1918 in support
of the Allied armies in their final concerted attack on the German armies along
the Western Front. This attack led to the end of World War One. The Royal Air
Force lost 96 aircraft on 8 August 1918 – forty-four to enemy
action and fifty-two to accidents that ‘wrecked’ the aircraft.
This ratio of accidental losses to losses due to enemy action did not elicit
any reference in the official record, so we may assume that it was not
particularly unusual to lose more aircraft to accidents than to the enemy.
Some seventy-five thousand members of Royal Air Force Bomber Command died in
World War Two. Once again, more than half of them died from accidental causes
rather than from enemy action. We can never know just what proportion of
these accidents had a root cause of ‘human error’ because human
error had not been closely defined then and accident investigation was, at
best, perfunctory.
The immediate post World War Two civilian accident statistics
reflected a similar situation. Airliners were either hastily repainted
military transport aircraft or newer transports developed from redundant
bombers. Improvements in aircraft design, in the reliability of all the
component parts of aircraft and in the training of those who operate aircraft
produced a steady reduction over the years in the accident rates for both
military and civilian flying but an analysis of the root causes of all
aircraft accidents highlights an area where there has been very little
improvement over the decades. As the number of accidents caused by mechanical
frailty reduced in number, the percentage caused by human error grew ever
more significant.
The More Things Change,
The More They Stay The Same
‘Human error’ was the root cause
of some 70% of aircraft accidents in the 1960s and human error remains the
root cause of about 70% of aircraft accidents today, forty years on. If we
can reduce human errors we should be able to reduce the accident rates in
just about every area of human endeavour.
There are two essentials that can lead to a reduction in human error
accidents. The first is to recognise and study just why these human error
accidents occur. The second, equally vital, essential is to encourage a free
and open system of reporting human error incidents in confidence so that we
may be forewarned of human error failures likely to lead to accidents. The
human error incident reporting programme must be confidential so that the
reporter knows that they are not exposing themselves to any kind of censure
from their hierarchy. The clearing house receiving the reports must protect
the confidentiality of each reporter so that the industry learns to trust the
system and to freely report their errors, and the observed errors of others. There
is no room for a ‘Blame Culture’ in any safety programme.
The United
Kingdom has a Confidential
Human Factors Incident Reporting Programme (CHIRP) that
encourages all participants in UK civil aviation to report any incident that
has a flavour of ‘human error.’ The reports received are
disidentified and then discussed at a quarterly meeting by a group who bring
a broad level of aviation experience to bear on each report. The reports are
then published, with comments, in “General Aviation CHIRP
Feedback” so that everyone is made aware of the human factors that led
to each reported incident.
Everyone involved in General Aviation, (pilots, instructors,
students, controllers and engineers) is encouraged to read each quarterly
issue of “General Aviation CHIRP Feedback” so that the importance
of human factors in Flight Safety is recognised. The importance of
understanding the nature of human factors and of human performance and
limitations cannot be over-emphasised. New pilots should be taught the basics
of human factors as part of their study for issue of their first licence.
This is a requirement of the International Civil Aviation Organisation
(ICAO). Pilots who have held their licences for many years may not have had
this formal training, but their need for an understanding of human factors is
equally important.
There are several sources of information on this subject. A
comprehensive range of titles on human factors and human error is available
from our
Bookshop. A
visit to the CHIRP web site at
www.chirp.co.uk will
provide information on the aims of CHIRP and how these aims are achieved. The
site also gives details of the individual who make up the CHIRP General
Aviation Advisory Board.
As Alexander Pope noted, “To err is human, to forgive,
divine.” We all make mistakes. When we make a mistake in aviation we,
hopefully, learn from that mistake. Why not share that learning with others?
There is no shame in making a mistake. Recognising our error and correcting
it can only enhance our safety. You can be pretty certain that others have
made the same mistake before you. There is a theory that each accident is
preceded by several hundred related incidents. If we can report these
incidents and recognise their causes then we can reduce the number and
frequency of human error accidents – ‘no mistake!’
Pushing the Limits
We
humans have been pushing the limits since we first got up onto our hind legs
and set off across the African plains to see what lay over the distant hills.
Curiosity is a great driving force, but my nanny often told me that
‘curiosity killed the cat.’ Many years passed before I began to
understand her warning. We need limits if we are to avoid stupid risks. We
haven’t time to discover all the risks in life for ourselves so we need
to learn from the mistakes of others, and avoid them, to survive.
Limits are
set in everyday life but our reaction to these limits varies from person to
person. It seems that it is human nature to push to, or beyond, any limits
set by others. This sometimes admirable human trait has expanded the
boundaries of our knowledge, but the same behaviour can lead us into
unexpected danger that is beyond our capability to handle.
Limitations in aviation are usually there for safety reasons. The
reasons for the imposition of these limitations may not always be obvious but
they are often based on hard-won experience. There seems to be a real problem
related to the perception and application of these limitations by some GA
pilots. Perhaps limitations are not well taught during the basic stages of
flying training. Perhaps their importance is gradually forgotten as pilots
gain flying hours and their self-confidence increases. The underlying problem
seems to be psychological in nature. Most car drivers recognise the fact that
if a vehicle strikes a pedestrian at 40 mph that pedestrian is much less
likely to survive than if they were involved in a slower collision. Yet many
drivers exceed the 30 mph limit in built-up areas.
Most drivers who do observe the limit seem intent on driving at
exactly 30 mph and revel in the gratifying sense of self-righteousness. The
safety message seems to become corrupted in their minds so that they think:
“The limit is 30 mph so if I drive at 30 mph I will be safe.” The
real message is: “It is particularly dangerous to exceed 30 mph.”
These two messages are quite different.
Pilots, being human, are also guilty of applying
faulty reasoning to the observation of aircraft limitations. If you tell the
average pilot that an aircraft has a Vne of 150 knots IAS then their mental
model is that they are safe if they keep the airspeed at or below 150 knots.
That’s not what Vne really means. I suspect that a fair proportion of
GA pilots do not know what Va is for their aircraft, nor fully understand the
implication of Va. If you tell them that their aircraft has a crosswind limitation of
16 knots they may think this means that it is safe and acceptable for them to
land with a crosswind component of 16 knots. That is NOT what the crosswind
limitation means.
A published cross-wind limitation (‘maximum demonstrated
cross-wind component’) means than a skilled and experienced test pilot
has demonstrated that he could maintain control of the aircraft at that wind
velocity – but not at one knot above that wind velocity. He will have
been in regular flying practice on type, in a fully serviceable aircraft at a
specific load and CofG position. The crosswind technique used on the
carefully chosen day will not be specified. But the limit was certainly not
determined on a short and narrow grass strip, surrounded by tall trees and
several large farm buildings, while granny and the kids offered advice from
the back seats.
Safety factors are applied to all aircraft limitations before they
are published. It is all too easy, particularly for inexperienced pilots, to
eat into these safety factors without intent. The point is that LIMITATIONS
are the outer boundaries of the permitted performance envelope and should not
be mentally distorted to become useable and acceptable numbers. Your own lack
of ability and currency may make it dangerous to even approach the specified
limitations of your aircraft. The results of exceeding the laid down limits
of your aircraft can be fatal, as described in the following accident
analysis.
Gruman AA-5B
A Grumman AA-5B was attempting to
land on Runway 08 at Luton Airport when directional control was lost. The
aircraft departed the runway and struck a parked and unoccupied turbo-prop
airliner. The three occupants of the Grumman AA-5B died.
There
are several important lessons to be learnt from this accident. Pick them out
as we follow the sequence of events that led to this fatal accident.
The front left seat occupant (FLSO) trained for a PPL over a 19-month
period and accrued almost 50 hours in that period. In the following 8 years
and 8 months before the accident he had accumulated another 92 hours, of
which 58 hours were as Pilot-in-Command. The purpose of the accident flight
was for the FLSO to gain time towards the minimum required within a 13-month
period so that his Certificate of Experience for his PPL could be
revalidated. He wanted to complete this before a change in regulations came
into force in four months time. His last Certificate of Experience was dated
11 months prior to the date of the accident and he had only flown three
flights totalling 1.7 hours as Pilot-in-Command since then. He had not flown
at all for 67 days prior to the accident.
The front right seat occupant (FRSO) had a total of 191 hours as
Pilot-in-Command with a further 156 hours flying experience. His most recent
Certificate of Experience was also dated 11 months prior to the accident,
since when he had flown 8.3 hours as Pilot-in-Command and 3.1 hours as P2,
all in the accident aircraft. He was a Licensed Aircraft Engineer and was
responsible for the maintenance of the Grumman AA-5B but did not hold any
Flying Instructor rating.
The right rear seat occupant (RRSO) had flown the accident aircraft
on at least nine occasions as handling pilot and had most flying experience
on board, but had not flown this aircraft during the past three years.
Prior to
the flight the FLSO checked the weather forecast for the morning. The surface
wind was forecast to be SSE at 11 kt, increasing between 0800 hrs and 1000
hrs to 20 kt with gusts to 34 kt. Luton Airport has a single asphalt runway
2160 metres long by 46 metres wide that is aligned 08/26. There is no
secondary runway. Thus the forecast crosswind for Runway 08 varied from a
minimum of 9kt to a potential maximum of 34 kt in the forecast gusts. A
checklist in the FLSO’s flight equipment listed the maximum allowable
crosswind component for a solo student pilot as 10kt and the ‘maximum
demonstrated’ crosswind for the aircraft as 16kt.
The aircraft was within the allowable centre
of gravity range and within the maximum allowable take-off weight limitation.
Taxi clearance was given at 0822 hr when the surface wind was 160/10 kt, a
crosswind component of 9.8 kt. Take-off clearance was given at 0831 hr when
the surface wind was 170/14 kt, a crosswind component of 14 kt. The takeoff
was unremarkable. The aircraft requested rejoin instructions at 0843 hr. It
was held on visual left base from 0848 until 0901 hr, awaiting a gap in commercial
jet traffic. The pilot of a Boeing 737 reported at 0850 hr that he had
experienced ‘a fair bit of chopping and airspeed change below 500
feet.’ The Grumman AA-5B was listening out on that frequency. The
Grumman was cleared to final and passed a surface wind of 170/14 kt, still a
crosswind component of 14 kt. It was cleared to land at 0902 hr and a surface
wind check of 170/17 kt was passed. The Luton Tower controller also held a
PPL and realised that the Grumman AA-5B might be affected by the crosswind
and by the turbulence on final. Further instantaneous wind checks were passed
as 170/20 kt and 160/20 kt as the Grumman crossed the threshold. Each
observation gave a crosswind component considerably (25%) in excess of the
‘maximum demonstrated crosswind component’ for landing a Grumman
AA-5B aircraft.
The Grumman flew along the runway at a height of about 10 feet with
the right, into wind, wing down and with some wing rocking. About 650 metres
into the runway the aircraft attitude changed to nose-up and then levelled
again. There was a marked left wing drop with the left wingtip touching the
runway surface before the mainwheels. The aircraft recovered to wings level
as it yawed 30° or 40° to the left (downwind) and left the runway to fly over
the grass before touching the grass surface. The controller initiated an
Aircraft Ground Incident and alerted the Airport Fire Service. The aircraft
appeared to accelerate and became momentarily airborne as it crossed the
parallel taxiway and then flew into a Shorts 330 correctly parked on the
south apron. The Airport Fire service was rapidly at the scene of impact,
some 1000 metres from the threshold of Runway 08.
The occupants of the right seats died instantly. The occupant of the
front left seat was unconscious when he was cut from the wreckage but he died
later in hospital.
The Grumman came to rest against the Shorts 330 with the
Grumman’s engine and propeller folded upwards at a right angle. The
rotating propeller had entered the cockpit. The flaps were UP at final
impact. The aircraft had flown at a low height above the runway for a long
distance before initial contact by the left wingtip with the runway. This
long distance may have been due to a high approach speed. The aircraft had a
stalling speed of 52 kt IAS at idle power with the flaps retracted. The final
impact against the Shorts 330 occurred with about 60 kt groundspeed. There
was a tailwind component of 13 kt at this point so it is unlikely that the
Grumman had sufficient airspeed to get airborne and climb before it struck
the parked aircraft.
Pilots and flying instructors are invited to carefully review the
facts and to list the sequential chain of errors that led to the deaths of
three men in a totally avoidable accident. Begin with the decision to get
into the aircraft before the flight and the self-induced pressures to do so.
Why did none of the three say, “Let’s forget it for today.”?
What were they thinking of while held in the visual orbit? Would you have
diverted to a more into wind runway – for example Runway 18 at
Cranfield?
The facts in this article are based closely on UK Air
Accidents Investigation Branch (AAIB) Field Investigation Reference
EW/C99/9/3 which source is gratefully acknowledged. The full report may be
seen by entering the above reference at www.aaib.gov.uk Any opinions
expressed are those of the author and are not intended to represent the views
of AAIB.
Text and Photographs © 2007 Gremline & Hill House
Publications, unless otherwise stated.
landing page
about gremline
copyright/conditions/contact
information exchange
glossary
uk emergency diversions
uk links, chirp & gasco
global & misc links
forum
the gremline cockpit — index of
articles
the gremline bookshop
top of page
