FIRE broke out on the Achille Lauro at about 10 pm on 30 November 1994. For
the elderly passengers enjoying a leisurely cruise in the Indian Ocean it was
the start of a hellish night. For the next seven hours the crew fought the fire.
Finally, at 5 am, the order came to abandon ship.
Although the sea was calm and the ship in no immediate danger of sinking, the
evacuation ran into trouble. Four passengers died: one was never found, one had
a heart attack and two died in the lifeboats鈥攊ncluding one person who was
hit by a life raft thrown from the ship. In the end, the Achille Lauro took two
days to sink. Things could have been much worse.
Cruise ships regularly catch fire, run aground or collide with other vessels,
and these risks are only likely to intensify. A few decades ago just
tens of thousands of people went on cruises but this year almost 10 million
people鈥攎ost of them from North America鈥攚ill holiday aboard ship.
This explosive growth has prompted a shipbuilding boom: 50 new liners will take
to the waves during the next five years and they will be bigger than ever.
Advertisement
But how safe will they be? Considering that the difference between a handful
of lives lost and a death toll in the hundreds may be only a matter of minutes,
you might presume that naval engineers had cracked the problems of getting
people off a ship fast. Nothing could be further from the truth.
Now a Canadian project is setting out to change all that. Engineers in
Ontario have built a giant ship simulator designed to rock and roll like the
real thing. This machine, they hope, will help them finally pin down the
problems that prevent passengers abandoning ship in a hurry.
The importance of evacuating a ship quickly has been recognised ever since
the Titanic went down in 1912. The International Convention for the Safety of
Life at Sea (SOLAS), drawn up in response to that disaster, stipulates that a
ship must be capable of evacuation in less than one hour. All emergency
planning鈥攖he position and number of lifeboats and muster stations, for
instance鈥攊s designed to comply with this rule. And to allow for an orderly
evacuation, ships are designed to stay afloat for at least an hour even after
they have been holed. Yet David Goodrich, president of the Royal Institution of
Naval Architects, admits that you鈥檇 never get everybody off in that time: 鈥淚t鈥檚
totally impossible. And we all know that it鈥檚 totally impossible.鈥�
The issue has been given added urgency by the rapid growth in the cruise
industry. Because of the mounting concern, Bill O鈥橬eil, the secretary-general of
the International Maritime Organization鈥攖he London-based UN agency
responsible for administering SOLAS鈥攍ast year ordered a two-year review of
the safety of large passenger vessels.
The problem the IMO faces is that nobody knows how long it really takes to
evacuate a ship. Many governments think that the best way to work it out is to
use computer simulations鈥攑articularly with models that reproduce the
behaviour of individuals, called 鈥渕icroscopic models鈥�. Microscopic models allow
individuals to move in opposite directions, the way real people move around
ships. But to be useful these models must be combined with real evidence from
evacuations.
Understandably, people don鈥檛 hang around with stopwatches when a real ship is
in trouble. Some countries have carried out trials, but these have been highly
artificial. In one of the largest trials, 842 people were evacuated from a ferry
in January 1996 in the calm waters of Dover Harbour. Everything favoured a
speedy result. The ferry had only half the normal number of passengers, and the
volunteers were mostly young and healthy. But it still took 90 precious minutes.
In a report last year, Britain鈥檚 Maritime and Coastguard Agency admitted: 鈥淭he
only real test is a genuine emergency with a high level of urgency which . . .
has not occurred in recent times.鈥�
Now the Canadians are looking for answers on dry land. In a suburb of Ottawa,
researchers are timing exactly how fast people can escape from the world鈥檚
largest ship simulator. Built by Fleet Technology of Kanata, Ontario, the
simulator was financed by the Canadian government after a small boat trial there
also failed to meet the one-hour rule.
A sinking ship will almost certainly be listing. It may also be pitching
around in mountainous seas, and some passengers may not be very mobile. So the
10-metre-long simulator has a corridor with a flight of 21 stairs at one end.
Hydraulic jacks can tilt the whole thing until it lists at 20 degrees. For the
first time, say the researchers, the tests will generate reliable information on
how people react when faced with the challenge of getting off a ship that isn鈥檛
on an even keel.
Fleet Technology has teamed up with Ed Galea at the University of Greenwich
in London to use these findings to make a realistic evacuation model. Galea had
developed software called Exodus for modelling the evacuation of buildings and
was modifying it for ships. A cruise ship, he says, is a bit like a
hotel鈥攂ut with crucial differences. 鈥淢ost passengers are very unfamiliar
with a ship鈥檚 structure,鈥� says Galea. 鈥淚n a building, you know you鈥檝e got to go
down to get out. On a ship you may have to go up or down to get to a muster
蝉迟补迟颈辞苍.鈥�
Women and children first
Tests in the simulator are designed to answer three key questions. How much
slower are old people than the youngsters used in most trials? On cruise liners
life jackets are kept in cabins. So a family rushing to get their life jackets
will have to cope with passengers coming the opposite way鈥攁lready wearing
their life jackets. How much does this hamper the evacuation? And how fast can
you get off a ship that鈥檚 listing badly?
So far, more than 200 volunteers have taken part. Each person is used only
once so they don鈥檛 learn how to escape more quickly. And the age range is
representative of the general public. 鈥淲e had all ages, from 4 to 81,鈥� says Ian
Glen, the president of Fleet Technology. 鈥淲e even had some folks with canes and
a wheelchair.鈥� This is vital, he says, because many cruise ship passengers are
pensioners. On the Achille Lauro, for example, the average age of the passengers
was 65.
Glen has tried a variety of tests. Individuals and groups have been asked to
walk along the corridor and up the stairs on an even keel and at varying angles
of list. At three points along the corridor they trigger optical sensors linked
to computers that automatically calculate how fast the 鈥減assengers鈥� move. Five
video cameras record the tests so researchers can study people鈥檚 behaviour in
detail.
To produce a useful simulation you need a decent length of corridor to get
accurate times, Glen adds. In the group tests you also need the space to see how
people behave in different circumstances鈥攕uch as teenagers hanging back to
help granny. There have also been tests with groups walking through the
simulator in opposite directions. 鈥淵ou should have seen the contraflow groups at
20 degrees,鈥� says Glen. 鈥淚t was quite a bunfight.鈥�
Some telling results have already surfaced. Below 10 degrees the list makes
no noticeable difference, says Glen. But beyond that the centre of the corridor
becomes virtually unusable and people lean on the walls for support. This
reduces the speed because people can get stuck behind others moving more slowly.
鈥淧eople will not overtake each other above a 10-degree list, especially on the
stairs鈥� says Glen.
With the 鈥渟hip鈥� tilting over, Galea expected everyone to walk along the lower
side of the corridor, leaning on the wall to stay upright. But the tests found
that a small but significant number of people chose to walk along the upper side
of the corridor, clinging onto the handrail. Neither Galea nor Glen is sure why
these people chose the more difficult option. But it may be a question of
weight鈥攖hin wiry people have less weight to support than the more
corpulent.
The tests with life jackets also surprised the researchers. Almost everyone
hesitated at the top of the stairs. 鈥淲ith hindsight,鈥� says Galea, 鈥渋t鈥檚
obvious.鈥� People couldn鈥檛 see their feet with the life jacket on so they stopped
because they were worried about falling.
Although the Canadian simulator has only been used to study list so far, it
can easily be adapted to simulate trim鈥攚hen the ship鈥檚 bow or stern is not
level鈥攐r the rolling of a ship in real weather.
Galea is also keen to test how people cope with smoke-filled corridors when
the ship is listing, to avoid any repeat of the Scandinavian Star disaster. The
ferry caught fire just after leaving Oslo in April 1990. At first the rescuers
thought everyone had got out. But two days later they found 158 bodies,
including 13 who had walked 3 metres past an exit.
Eventually he hopes his maritime version of Exodus will pinpoint the
bottlenecks in an evacuation. Then it should be quite easy to minimise hold-ups
by planning evacuation routes that skirt round them, or even redesign the ship鈥檚
layout.
So when are these lessons likely to be put into practice? Unfortunately,
changing existing safety rules will take some time. Any IMO decision on safety
depends on complex international negotiations, and some in the shipping industry
may lobby against new rules that could reduce profit.
One of the more likely outcomes of the IMO鈥檚 current review is the
introduction of dual certificates for ships. At present, liners are certified to
carry a maximum number of passengers, regardless of people鈥檚 age. But the
researchers believe that ships should be forced to carry fewer passengers if
they are all elderly鈥攂ecause of the extra time it takes to evacuate
them.
According to IMO spokeswoman Natasha Brown, committees are scrutinising
evacuation times. If the one-hour rule is broken, they will make proposals to
achieve it. 鈥淵ou should give us credit,鈥� she adds. 鈥淲e are looking at the
颈蝉蝉耻别蝉.鈥�
Almost seven years after the Achille Lauro fire, the shipping industry
finally seems ready to confront an uncomfortable truth. Cruise ships simply
can鈥檛 be evacuated in the one-hour limit, and something must be done before a
huge disaster hits the headlines.