Children are being killed on Britain’s roads because of an optical illusion,
warn researchers at the University of Aberdeen. The illusion fools drivers
into underestimating how long it will take them to stop, and by the time
they discover their mistake it is too late to avoid an accident.

About 50 children are injured on Britain’s roads every day. Investigators
looking at this problem have tended to blame children’s inexperience and
carelessness. But the researchers at Aberdeen suggest that the problem lies
with the drivers. Doug Stewart of the university’s engineering department
says that the optical illusion is responsible for more than half the injuries
to children on the roads.

People normally judge the time it will take an approaching object to
reach them by the rate at which its size increases, a phenomenon called
optic flow. This is adequate for catching a ball or ducking a projectile,
Stewart says, but for car drivers it is only effective in the final second
or so before collision. At speeds of 50 kilometres an hour, which are typical
for traffic in towns, it takes around 3 seconds to stop a car.

Stewart and his colleagues Chris Cudworth and Roly Lishman say that
instead drivers judge the time before a collision by subconsciously measuring
the angle between the top and the bottom of the pedestrian’s image (see
Diagram). In the absence of any other clue they assume that the pedestrian
is an adult and judge the distance accordingly.

If the pedestrian turns out to be a child this assumption can be lethal,
as it leads drivers to believe the child is farther away than it really
is. By the time the error has become apparent it can be too late for the
driver to avoid hitting the child.

Stewart and his colleagues conducted a series of tests in which volunteers
acted as drivers in a computer simulation of a car approaching a pedestrian.
In the test, volunteers were presented with a child 1 metre tall and an
adult 1.8 metres tall. The volunteer ‘drivers’ were asked to stop a safe
distance in front of the pedestrian.

Describing their results in the journal Perception, the researchers
say that drivers were more likely to make errors when the image was a child’s.
They also found that the driver’s height above the road influenced the number
of errors. A driver sitting 1.8 metres above the road was less likely to
misjudge the distance than one sitting 1 metre above the road.

This seems to be borne out by the road accident statistics. Cars are
two to three times more likely than lorries to hit children, when the number
of each type of vehicle on the road is taken into account. Lorry drivers
sit higher up than car drivers.

Stewart says that the risk of drivers making potentially fatal errors
can be reduced by providing visual cues for drivers about the size of the
object they are approaching. Britain has one of the worst child safety records
in Europe. One reason, says Stewart, is that other European countries have
more zebra crossings, and that the markings on the road help drivers to
distinguish between a child and an adult.

Britain’s zebra crossings are gradually being replaced by pelican crossings,
which have traffic lights but no striped markings on the road. Stewart says
that pelican crossings should also have road markings to help drivers judge
the size of a pedestrian.

In correspondence with Stewart, the Department of Transport describes
the research as ‘interesting’, though it does not accept that the paper
‘demonstrates that perceptual error in the way described is the main reason
for children having a much higher accident rate’. The DoT does agree, however,
that marking the road with stripes at pelican crossings could improve things.
Unfortunately, says Stewart, there is no indication that the positive tone
of some of the DoT’s comments has led to ‘action to curb child pedestrian
accidents or to promote further research’.

There is an uncomfortably close parallel between the development of
roll-on roll-off ferries – a classic case of safety rules lagging behind
technology and design – and the development of high-speed catamarans over
the past decade. In the case of ro-ro ferries the legislation only began
to catch up after the Zeebrugge tragedy of 1987 in which 193 people died.

New international regulations now under discussion will make this test
– a thorough theoretical survey of everything that could go wrong and its
effect – compulsory, but not for at least three years. The biggest catamarans
are 74 metres long, and can carry 400 passengers. Plans are on the drawing
board for 110-metre long craft, capable of carrying 1400 passengers and
200 vehicles.

This week an Australian-designed high-speed catamaran began a new service
across the Channel. A second is due to enter service later this month. They
will replace Hoverspeed’s ageing fleet of hovercraft, the technology of
the 1960s. Ironically this vital safety test, known as failure mode and
effect analysis, was carried out on the hovercraft.

The concern about high-speed catamarans first surfaced following a serious
accident in Hong Kong at the end of 1989. This accident has barely been
reported outside the colony. Earlier this year the Hong Kong government
submitted extracts from the inquiry report to the International Maritime
Organization, the agency which is responsible for administering conventions
on the safety of life at sea. As a result of the accident both the colony
and the British government have been pressing the IMO for urgent action.

On the evening of 15 December 1989 the Swedish-built Apollo Jet catamaran
had just unloaded its passengers at the ferry terminal. It set off for its
overnight moorings at its full speed of 30 knots.

The crew suddenly lost control of both steering and propulsion. The
Apollo Jet veered sharply to starboard at full speed. In the next 71 seconds
the catamaran careered across the crowded harbour, running down a sampan
and a passenger launch and colliding with a number of other vessels, before
climbing the sea wall. It ended its voyage in a typhoon shelter. Six vessels
were badly damaged. Four people died.

A marine court investigated the accident in October and November last
year. It found that the cause of the accident was a single switch in the
24-volt control system.

The court concluded: ‘In simple terms Apollo Jet . . . is a vessel that
relies for its safety and control on its 24 volts system.’ Although the
designers understood the importance of backup systems – the control system
had no fewer than six separate sources of power supply – they destroyed
this advantage by merging these sources in a single electrical circuit with
a switch which was ‘open to abuse and maloperation’.

‘There was . . . no proper backup system to the most important supply
in the ship. In the opinion of the court this design weakness or failure
was so serious and fundamental as to make the vessel unseaworthy.’ The court
said the fault should have been ‘seen and corrected’ by the vessel’s designers,
builders, the Hong Kong marine department and the society, Det Norske Veritas,
which surveyed the ship.

However, the court highlighted a more fundamental reason for the accident.
Although the court had ‘no hesitation in finding that’ these craft were
‘as safe as any other . . . It might be asked therefore why the accident
to Apollo Jet occurred. The answer is simply that legislation, management,
surveying, training and manning have not kept up with the demands imposed
by design.’ The court said that these craft were more akin to aircraft
than ships. ‘So far as control is concerned Apollo Jet more closely resembles
an aircraft than a conventional ship.’ Yet the design principles date from
the days of ocean liners.

‘The conventional ocean-going passenger ship is designed to go to sea
independently for an extensive period . . . . Generally speaking, emergencies
are floods or fires or approaching vessels, all of which can be dealt with
in a relatively slow time. There is very little on board a passenger ship
in the open sea that requires immediate reflex action . . . ‘ The court
says that dynamically supported craft such as hydrofoils and catamarans
‘are of a totally different concept and design. They are required to go
on short journeys close to land at high speed . . . ‘.

The court sees ‘nothing wrong with . . . aircraft control features within
a maritime craft’ but it says that designers have not been sufficiently
aware of the dangers of ‘relying on sophisticated control systems which
do not have the built-in reliability and redundancy that such remote control
systems demand . . . ‘ The report’s key recommendation is that a failure
mode and effect analysis should be carried out on every type and class of
these high-speed craft. This analysis should ‘confirm that there are at
least two and preferably three totally separate, independent and guaranteed
supplies to the control circuitry for lift, direction and propulsion.’ In
the event of complete failure the system should fail safe. The recommendation
has added force because Hong Kong has wide experience of high-speed craft.
The colony has the greatest concentration of high-speed vessels in the world,
a total of 60 catamarans and hydrofoils.

Analyses of these sort and the closely related technique of fault tree
analysis are common practice in aviation and the offshore oil industry.
The Civil Aviation Authority says this analysis is carried out ‘in both
directions along the tree’ on all passenger aircraft. The authority said
that hovercraft were ‘no different at all from aircraft in their safety
³¦±ð°ù³Ù¾±´Ú¾±³¦²¹³Ù±ð’.

Marshall Meek, president of the Royal Institution of Naval Architects
said: ‘If you introduce new systems you must adopt modern practices.’ Kenneth
Rawson, who chairs the Royal Institution of Naval Architects’ committee
on ship safety said ship design practices were archaic. Fault tree analysis
was ‘absolutely essential to any major engineering project. We have to move
into this century’.

Gerry Dodd, of British Maritime Technology, who is an expert on risk
analysis, said that ‘until you do these analyses you cannot be sure you
have eliminated the problems’.

The design of ships is governed by an international convention on the
safety of life at sea, known by its acronym SOLAS. However, catamarans come
under a separate IMO code for dynamically supported craft, which was agreed
in the early 1970s. This was intended to lay down safety standards for hydrofoils
that were equivalent to the SOLAS convention of 1960.

Since the code was produced the SOLAS rules have been tightened up considerably,
notably in 1974 and following the Zeebrugge disaster. The code for dynamically
supported craft remains unchanged. And high-speed catamarans had not been
developed when the code was drawn up.

In May, Britain and Hong Kong persuaded the IMO to accept failure mode
and effect analysis as the philosophical cornerstone of a new code, which
will apply to all high-speed craft that travel faster than 25 knots. The
code is now being drafted. Some countries, notably Australia, have reservations
about the new code. If all goes well, it should be completed in time for
approval by the IMO’s general council at its meeting in November 1993 and
come into force the following year.

Rawson criticised the slowness of international negotiations. He said
Britain was ‘tied to an international system which is so deadly cumbersome
people can hide behind it’. New ÐÓ°ÉÔ­´´ understands that Britain may require
catamarans using its ports to have an analysis before agreement has been
reached in the IMO. However, it will take at least a year for the government
to develop an analysis.

Claus Norrstrand of Marinteknik Verkstads, the firm which made the Apollo
Jet, said it would carry out failure mode and effect analysis when international
rules required it. He said there had been ‘some discussion with classification
societies but no one has been able to tell us how it should be carried out’.

Governments delegate part of the responsibility for safety to the five
major classification societies, such as Det Norske Veritas and Lloyd’s Register.
These societies produce their own rules and survey vessels. Without approval
from one of these societies a vessel is virtually uninsurable.

Det Norske Veritas classifies 80 per cent of the world’s high-speed
catamarans. The society revised its rules for these vessels in January but
the new rules do not include a failure mode and effect analysis on the control
systems of a vessel. The society said of the analysis: ‘We believe the concept
is a good one, and is within the spirit of the IMO code. We feel, however,
that it needs further development and should be more specific.’ International
Catamarans of Tasmania is the firm that developed the wave-piercing catamarans
which began running across the Channel this week. The craft have had a troubled
debut both on the Channel routes and in Australia. On the channel run last
year passengers had widely publicised problems with seasickness (This Week,
10 November 1990). The Bass Strait ferry between Australia and Tasmania
had to be taken out of service because of cracks in a British-built engine.
It is due to start again later this month.

The company’s corporate objectives are ‘to establish a streamlined boat-building
facility that is smaller but on similar lines to that of large aircraft
production lines’.

In Tasmania Gary Linacre, project coordinator for the company, said
he did not know what was meant by failure mode and effect analysis: ‘Extensive
sea trials have been carried out.’

Britain’s Changing Environment by Tim Bayliss-Smith and Susan Owens
(Cambridge University Press, pp 252, 25 Pounds is a fascinating colection
of aerial photographs of the British countryside. From motorways to power
stations and the eponymous white cliffs, the photographs reveal people’s
innate ingenuity in finding ways to disfigure the countryside.

But this is more than a catalogue of warts. The photographs show the
crop marks of old river meanders and prehistoric settlements, while the
flying lens graphically captures the immense natural challenge of coastal
erosion. Living in towns such as Peacehaven is literally a cliffhanger,
with whole streets poised to vanish.

The first TGV line, between Paris and Lyons, opened in 1981. The second
TGV line, the TGV Atlantique which connects Paris to the west coast, opened
in part in September last year. The third TGV line, the TGV Nord, which
will link Paris to the Channel Tunnel, Brussels and Cologne is under construction.

The French and West German railways have two differing technologies
for their high speed trains. The West German ICE (InterCity Experimental)
took the world speed record in 1988 with a run of 407 kilometres per hour.
Late last year the French TGV regained the record with a run of 482 km/h
on its TGV-Atlantique line between Paris and Tours.

West Germany’s two new high-speed lines will begin operation in June
1991. They will carry the ICE train, which is similar in shape to the French
TGV and carries a similar three-phase electric motor. Trains on the line
will average 250 kilometres per hour. Because the distances between cities
in West Germany are shorter than in France, says Hans Hauschild, of West
German Railways, higher speeds do not make economic sense. West Germany
plans 2000 kilometres of high- speed line over the next 10 years.

The first moves towards making these technologies more compatible –
and allowing the trains to cross frontiers – began last year with a test
run by the ICE over French rails. The biggest problem was thought to be
that the TGV has a maximum axle load of 17 tonnes. The ICE with its higher
axle load would be too damaging for French track.

The test run proved that these fears were overstated, but in November
the agreement on the TGV Nord between Belgium, France and West Germany stipulated
heavier rails for this line.

The only remaining difficulty in linking ICE and TGV lines, says Hauschild,
is that the ICE is wider, and provides more comfort than the TGV. The new
French TGV-Atlantique is more comfortable than the first version, says Hauschild,
but discussions are now under way between the two railways. The French are
worried about consumers’ preference, since both trains will have to charge
the same price for the same journey.

Nonetheless, differences do remain. The French high-speed lines carry
passengers only. German lines also carry freight. The German trains cannot
handle steep gradients as easily as the TGV. So the West German lines have
more expensive bridges and tunnels.

Europe’s future high-speed rail network is slowly taking shape. The
French, Germans, Dutch and Belgians have agreed on a set of links due to
be complete by 1998. Most of these links will receive generous government
subsidies. The British government has so far refused to subsidise high-speed
trains, despite persistent rumours that up to Pounds sterling 1 billion
could be forthcoming for the link from London to the Channel tunnel.

A further factor that will influence the development of high-speed rail
links is the unification of Germany. A major programme to strengthen conventional
rail links to Eastern Europe has been launched by the UN Economic Commission
for Europe.

Meanwhile, Germany’s Transrapid, a ‘magnetic-levitation’, or maglev,
train has been put on hold. The West German government decided in December
that a successor to the Transrapid test track in northern Germany is to
be built between the airports of Cologne and Dusseldorf.

The Institution of Professionals, Managers and Specialists, the union
that represents the majority of scientists, reacted angrily. ‘These plans
have been shrouded in secrecy,’ says Valerie Ellis, the union’s assistant
general secretary. She expressed concern that as a result of cutbacks in
research staff, the museum would shift from long-term scientific excellence
to catering for the latest fashions.

The museum has already cut 50 posts since 1984. The details of the new
cuts are spelt out in the museum’s corporate plan, which puts the cost of
the redundancies at Pounds sterling 1.8 million. This is roughly what the
museum receives from admission charges, which earlier this year rose from
Pounds sterling 2 to Pounds sterling 3 a head.

According to the corporate plan, a copy of which has been seen by New
ÐÓ°ÉÔ­´´, the number of scientists employed by the museum will fall from
299 to 253.

The cuts fall across the board. The 49-strong botany department is to
lose 10 posts, research on fossil birds and plants will go entirely, and
palaeontology will lose 11 of its 54 staff members.

The corporate plan envisages a refocusing of the museum’s activities.
It specifies that in future the museum’s scientific expertise will be concentrated
in six areas: biodiversity, environmental quality, living resources, mineral
resources, human health and human origins.

The museum’s director Neil Chalmers argues that although nobody wants
to reduce staff; it was preferable to have planned reductions rather than
haphazard cuts. ‘We need new blood on short-term appointments in particular
areas, in fast moving subjects,’ Chalmers said on Monday.

The museum’s corporate plan opens by arguing that increasing concern
about the natural environment makes the museum’s expertise more relevant
than ever. ‘These new and expanding demands require an organisation which
is . . . soundly housed and properly resourced,’ it says.

However, the reduced grants requires the museum to ‘cut its cost according
to its cloth’. This necessitates a reappraisal of the priority areas for
science and a closer definition of posts and the management structure, says
the plan.

The net effect is ‘to reduce the existing staffing levels by 50 to 60
posts, in addition to the 40 posts which have been saved in 1989-90 through
natural wastage’.

In 1986 the corporate plan said that redundancies would be implemented
‘only if totally unavoidable’. The current corporate plan says the reduction
will be achieved ‘by natural wastage and redeployment where possible but
some voluntary and compulsory early retirement will be necessary’.

The unions fear that more cuts could be in the offing. The corporate
plan predicts a Pounds sterling 500 000 shortfall, even after the job cuts,
which it hopes the government will fund. The root of the museum’s financial
problems is that the government is increasing its funding by 4 per cent
annually. The inflation rate assumed in the plan is 7.5 per cent.

Union sources said they had asked for a meeting with management when
financial problems first became evident late last year. Chalmers met the
unions to discuss the corporate plan cuts on Monday this week. Before the
meeting, he said the corporate plan provided a sound basis for the future
of the museum.

The inquiry heard that the most likely cause of the disaster was a poorly
fitted flange – a cap for a pipe – which allowed gas to escape and explode.
However, no surviving eyewitness saw the initial sequence of events. In
addition, the centre of the platform still lies 150 metres under the sea
beneath some 4000 tonnes of tangled metal.

On the afternoon of the disaster, maintenance engineers had taken a
pump out of operation for service. At the same time, the engineers took
the opportunity to withdraw a pressure relief valve from the line and service
it. The normal procedure was to plug the line temporarily with a flange.
They put the pump back on the pipe, but could not finish work on the pressure
relief valve. So they left the temporary flange on in place of the relief
valve.

Some 20 minutes before the disaster, another pump broke down on a parallel
line, which at the time was coping with all the platform’s workload. Engineers
failed to restart that pump, and tried instead to put the newly reserviced
pump back into operation.

The gas detectors set off alarms at about 2156 hours, showing that locally
there was an airborne concentration of about 15 per cent of an explosive
mixture. At 2200 hours, a huge explosion shook the oil rig. The investigators
were able to establish the time to within 20 or 30 seconds because the crew
were watching television news as the blast struck.

The explosions could have been caused either by a mixture of ethane
and methane, or by condensate, a mixture of slightly heavier paraffins and
water vapour. Pipes carrying both gases passed close to the centre of the
explosion. British Maritime Technology conducted wind tunnel tests – the
wind speeds through the platform were known reasonably accurately because
of data from weather monitoring ships nearby – which demonstrated where
the activated gas alarms were situated. They were all close to floor level,
indicating that the gases were heavier than air.

In addition, tests on models of the Piper Alpha structure showed that
the explosive force could have been created by between 150 and 225 kilograms
of methane and ethane, or 45 to 75 kilograms of condensate. The wind would
probably have dispersed the methane and ethane.

The investigators concluded that the leak had to be condensate, and
that the source was most likely to be the flange. Further tests on the physical
distribution of the gas alarms that went off showed that the gas must have
come from an opening equivalent to a hole 10 millimetres across, and in
the form of either a jet or a fan, with a leakage rate of about 100 to 120
kilograms per minute. The most likely occurrence that matches this scenario
was a leak from a flange that had not been properly tightened.

However, another possible cause of the explosion is the build-up of
hydrates, or water trapped in crystalline structures, which blocked a pipe.
This could have caused a build-up of liquid in the compressor which would
have eventually led to an explosion, by blowing the pipe apart.

The Department of Energy was heavily criticised at the inquiry report,
especially by the trades unions. About half a dozen people died in the original
explosion. The department, as a matter of policy, limited the number of
pipelines on the bottom of the sea. So three satellite rigs passed their
oil and gas to land through Piper Alpha. Their communications were also
centred on Piper Alpha.

All the satellites lost communication with Piper Alpha when the rig
exploded, although some managed to maintain contact with the shore. In addition,
none of the satellites could flare off hydrocarbons quickly enough in an
emergency to relieve the pressure in the pipelines. Consequently, the pressure
in the pipe meant that oil and gas continued to feed the fire on Piper Alpha
for up to an hour after the original explosion.

Piper Alpha accommodated about 800 tonnes of flammable material. Altogether,
the three satellite rigs and Piper Alpha accommodated about 2400 tonnes.