¿ìè¶ÌÊÓƵ

British Midland Airways last month sacked the two pilots who were at
the controls when a Boeing 737-400 smashed into the central reservation
of the M1 motorway in January 1989. The crash killed 47 people.

The report into the Kegworth disaster*, published last month, states
that the pilots acted hastily and contrary to their training. Yet it also
points out mitigating circumstances for their actions, highlighting the
fact that pilots operate in a complex and changing technical environment.
A pilot’s actions might be but one part of a set of circumstances that leads
to an accident.

Some of the observations about the environment in which the two pilots
were working also highlight a general area of concern for safety in modern
aviation: the interface between pilot and machine and how it affects a pilot’s
performance. Frank Taylor, from the Cranfield College of Aeronautics, says:
‘If a pilot makes an error, it’s often because he or she has fallen into
a hole that someone else has dug for them.’

Regulators and academics in the civil aviation world, though lagging
considerably behind their military colleagues, are beginning to recognise
the impact on safety of what are called ‘human factors’. The relationship
between machines and people is one aspect of the topic.

The evaluation of human factors in the cockpit is taking place against
a background of increasing automation. The aviation world has dubbed those
aircraft with highly automated flight decks ‘glass cockpits’. The name comes
from the glass TV screens that take the place of the devices which previously
displayed flight infomation.

The electronics and computer technology behind these glass screens mean
that information can be presented in different ways. It means, too, that
machines can undertake tasks previously executed by pilots. Don Harris,
from the applied psychology unit of the Cranfield College of Aeronautics,
says: ‘Pilots are changing from fliers to flight managers.’

In many cases, the new technology leads to greater safety. For example,
pilots wax lyrical about the benefits of what is termed the map mode of
the horizontal situation indicator (HSI). This instrument combines the essential
information about the plane’s lateral and vertical position and speed with
respect to the ground and superimposes the information on a map.

In the early days of aviation, pilots would struggle with raw numbers
and a separate map. Alistair Liddle, a pilot and head of the technical committees
of the British Airline Pilots’ Association, calls the instrument ‘the best
thing since sliced bread. It combines important information well, and pilots
can easily assimilate what it says.’

Yet not all displays are as well received by pilots. And the aviation
world needs to query whether the new displays present information to the
pilot as clearly as traditional instruments do. The Kegworth report emphasises
this issue.

Soon after take off, last January, the crew experienced ‘moderate to
severe vibration’ and smelt smoke. The co-pilot monitored the instruments
showing the health of the engines. When the commander asked which engine
was causing a problem, the co-pilot started to say the left then switched
to say the right engine.

He obeyed the commander’s order to throttle back the engine, and as
he throttled back the right-hand engine, he clearly still thought that this
was the faulty engine. Tragically, he was wrong. As the plane came in to
land, the operating engine failed, and it was too late to restart the good
engine.

Analysis of the flight data recorder and of the wiring to the engine
instruments in the cockpit proved that the instruments would have shown
accurately which engine was suffering from unusual vibration. When questioned
later, the co-pilot could not remember what he saw on the instruments that
convinced him to switch off what turned out to be the wrong engine.

The report speculates that the hesitation between left and right could
have been caused by genuine difficulty in reading the instruments. The investigators
point out that the pointer on the engine vibration monitor was far shorter
than its electromechanical counterpart. Alternatively, the co-pilot may
have have observed the instruments during a six-second period when, according
to the flight data recorder, the instruments for the left hand engine showed
‘relative stability’.

In a later incident, a Dan Air crew who were flying a Boeing 737-400
experienced similar conditions of vibration and smoke. The commander identified
the faulty engine correctly from readings on the engine instruments. But
the pilot told investigators that all the engine instruments looked alike,
and that it took time for him to be certain that he had identified the affected
engine. Fortunately, he had the cautionary tale of Kegworth to guide him.

One of the recommendations from the investigation into Kegworth is that
the rules for certifying new instruments should be modified to include standard
tests for assessing how well instruments relay information to the crew in
both normal and abnormal conditions. The investigators recommend that airline
pilots should be involved in the process – a view Liddle agrees with.

Currently, test pilots and senior pilots, not the pilots flying routinely
for airlines, are asked to assess the effectiveness of instruments. One
former flying instructor says that test pilots are specialists at the top
of their profession and that they cannot be expected to react in the same
way as the airline pilots. Liddle, an airline pilot, says he has observed
in simulator training that test pilots and senior instructors react very
differently to the airline pilots. ‘Perhaps airline pilots are more practical,’
he says.

Computer technology allows designers to call up on screen alternative
instrument designs relatively easily. The time it takes an airline pilot
to read information from the simulated instruments and the number of errors
the pilot makes would provide objective data on the effectiveness of a particular
design. And the tests need to take place when the pilot is under stress,
not only during routine operations.

The Kegworth report’s criticism that the pilots acted contrary to their
training is based on the fact that they acted hastily. Yet the report points
out that neither pilot had experienced a combination of high vibration and
smoke in the cockpit, either in training or in operation. The report says
the combination of circumstances appears to have driven them to act quickly
in response to what they perceived as a serious engine malfunction with
an associated fire.

Neither pilot had practised fault diagnosis with information from those
instruments. In fact, in the past, pilots often disregarded readings from
vibration monitors because they believed them to be unreliable. After the
Kegworth accident, the pilot, Captain Hunt, expressed this view.

However, advances in technology mean that monitors, such as those on
the 737-400 are far more reliable, and the accident investigators recommended
that pilots should be made aware of technical improvements that have turned
what was a rarely used instrument into one of more significance. The Civil
Aviation Authority has accepted both this recommendation and another that
advocates using simulators to teach pilots how to interpret information
displayed by the vibration monitors.

The question of training and automation, though, is much broader. Currently,
pilots tend to be taught to read what the instruments are saying, rather
than how the system works. One aviation psychologist believes that this
needs examining.

Clearly, he says, pilots do not need to know how to program the machines,
but they should know a bit about how they work. Comments heard on the flight
deck include, ‘I’ve never seen it do that before’ or ‘Why’s it doing that?’
Anyone who has operated a word processor must have said the same thing at
some time. A pilot may not have time to work out why his system is throwing
up unfamiliar signs.

The specific questions raised by the accident investigators are to be
reviewed by the accident investigation and the air worthiness committees
of the British Airline Pilots’ Association, says Liddle. At the same time,
Roger Green, head of psychology at the RAF’s Institute of Aviation Medicine,
is conducting a survey of pilots’ views of automation. Green, who devised
much of the syllabus for the CAA’s new pilots’ examination on human performance,
is currently analysing responses from 1300 pilots.

A first glance at the responses shows that many pilots are well satisfied
with automation, but that there are important reservations. Some of the
older pilots worry that sophisticated technology can give young pilots a
false impression of their flying abilities. When they return to an aircraft
with more traditional technology, they have to unlearn bad habits.

Other pilots note that when everything is going well during a flight,
automation works well and reduces the pilot’s workload. If a problem arises,
automation can increase the workload at the worst possible time. At Kegworth,
for example, the co-pilot had to reprogram the flight management system
with new information about diverting to east Midlands Airport. It took two
minutes to do this at a critical time in the flight.

Crucial exchanges with air traffic control and attempts to reprogram
the flight management system meant that the pilots never had time to complete
a review of their decision and their response to it, although the Commander
did initiate this.

Earl Wiener, from the University of Miami, has studied the benefits
and hazards of cockpit automation. He believes that when new automated cockpits
are certified, regulatory authorities should assess the workload during
abnormal circumstances, not just when everything is going well.

Yet another concern expressed by British pilots centres on what they
term ‘mode awareness’. ‘They may think,’ says Green, ‘that the aircraft
is programmed to level out at a particular altitude, but it isn’t, so they
overshoot, and may have a near air miss.’

All of these concerns demonstrate that a pilot can be responding to
an unusual situation in what they eive to be the optimum way. Yet, in the
changing high-technology environment, they may be tragically wrong. Perhaps
an instrument they disregarded as unreliable has become more reliable; perhaps
some combination of design and lack of training contributes to wrong decisions.

At a meeting in London this week, Wiener told delegates at a conference
on safety in the air and at sea that technology and safety must be placed
in the broad context of the environment that a system works in. He warns:
‘There is a tendency to blame accidents solely on the last person in the
chain of events, usually the captain.’

* * *

Health and fitness provide key to safety

From next year, the Civil Aviation Authority will insist that candidates
for a pilot’s licence sit an examination entitled Human Performance and
Limitation. This is designed to make them aware of their limtis and help
them to understand their relationship with technology and with one another
on the flight deck.

In the first part of the course, pilots will learn some basic physiology,
and an understanding of what happens to the body at high altitudes and how
to counteract the effects. Another module on flying and health will teach
them about their own health and how to stay fit enough to pass their six-monthly
medical examinations.

Basic aviation psychology will teach them how people process information,
and how their own perception of the world can be unreliable. Pilots who
land short, for example, may have misperceived the physical cues intended
to help them in their approach to a runway.

Stress and sleep management also forms part of the course.

‘Pilots are particularly vulnerable to sleep problems. Their circadian
rhythms are mucked up by the long haul flights,’ says Roger Green, of the
RAF’s Institute of Aviation Medicine at Farnborough.

Finally, pilots will be introduced to ergonomics and elementary social
psychology. How, for example, should a leader behave?

‘First,’ says Green, ‘a commander should listen to his or her crew,
rather than putting forward their own ideas.’ that way, they do not discourage
a junior colleague from expressing a view that may contradict the commander,
yet could be vital to dealing correctly with an emergency.

*Report on the accident of the Boeing 737-400 G-OBME near Kegworth,
Leicestershire, on 8 January 1989. Produced by the Air Accident Investigation
Branch of the Department of Transport. Available from HMSO.