This article explores how HSL can help establish the reasons
behind major incidents.
If a picture is worth a thousand words the laser scans produced
by the Health and Safety Laboratory (HSL) must be worth several
billion. This is especially so when their purpose is to help
determine the causes of major disasters.
"The first step is usually a telephone call from the HSE
principal inspector at the scene of the incident," says Steve
Graham, head of HSL's Visual Presentation Services Unit. "We
have to move quickly to get a photographer to the scene."
"Photographer" is perhaps misleading. It brings to mind
the silver halide cameras used by Steve Graham when he first joined
HSL. HSL's use of imaging is way beyond that. Their
pioneering use of laser scanners is just one example of the ways in
which HSL puts cutting-edge science into accident investigation.
Laser scans taken at different points in the accident scene are
combined to produce "wire frame" 3-D image of the
"We can also put texture onto this model to produce something
that looks more like a conventional picture," says Steve.
Of course it is more than just a "picture". This
becomes obvious as one of Steve's team demonstrates its power on
his computer screen. The way he zooms around the scene is
reminiscent of "Minecraft" but with infinitely superior graphics.
He can use the computer model to generate pictures from any
point and angle, something that would take, quite literally, an
infinite number of conventional digital photographs.
Nor could digital photos be used to provide an accurate measure
of the distance between any two points at the scene of the
"Cheaper, quicker and less labour intensive than using
theodolites," says Steve Graham. "And if we subsequently need a
measurement we did not make a note of at the time we can simply
take it from the model.
HSL can apply data from other disciplines to the scans to help
elucidate the cause of the incident. If the incident results
in a court case they can also rotate and animate the image to give
judges, barristers and jurors a clear picture of exactly what
But this does not mean HSL have consigned the more conventional
imaging techniques to the garbage heap of technological
history. These still have their uses. For example they
are light enough to mount in drones and beam back pictures
from, say, a burning factory.
"From imaging data we can often get a good idea of how the
incident started," says Steve Graham
Chris Keen is a principal occupational hygienist. He is
based at HSL's Buxton site deep in the Pennines. Most of his
team however are located in HSE offices around the UK where they
work with specialist HSE Inspectors.
"We can thus get occupational hygienists to the site of an
incident very quickly," says Chris.
He identifies three main incident response services provided by
"Firstly we collect samples from the site. These can
include substances such as asbestos, chemicals present initially
and chemicals produced by the incident. Now we can use drones
to collect airborne chemicals."
"Secondly, we collect articles. For example we may recover
pieces of ruptured tank scattered across the site by an
explosion. We may also recover pieces of machinery whose
failure may have resulted in the incident."
But their role is not confined to sample retrieval. Their
third function is to help elucidate causes. Even today, all too
many incidents investigated by Chris's team arise from work in
"The majority of these fall into three groups," he explains,
"oxygen depletion (the largest), carbon monoxide poisoning and
hydrogen sulphide poisoning.
Sometimes the cause of death is not immediately obvious.
"We investigated a triple fatality that occurred on a
ship. The bodies were found in the ship's anchor chain
locker. Not an obvious high risk confined space. Our first
reaction was 'what could possibly be the risk in that kind of
environment?' The culprit actually turned out to be the
anchor chain. It had been hauled from the sea dripping with
salt water and left in the locker room for days on end.
During that time it had rusted. Rusting is a chemical reaction that
takes oxygen from the air. As a result oxygen levels in the
locker room had reduced to, as it turned out, a fatally low
Sometimes the oxygen depletion results from displacement by
inert gases such as carbon dioxide:
"A worker accidentally dropped his mobile phone into a tank
containing pig feed. He climbed down a ladder into the tank to
retrieve it. He was overcome and died. It turned out
that the pig feed consisted of fruit cake supplied by a local
baker. The feed had fermented to produce an atmosphere of
carbon dioxide at the bottom of the tank."
Tragically, confined spaces all too frequently lead to multiple
deaths when fellow workers rush to help a collapsed colleague and
are themselves overcome.
Traditional instruments such as electrochemical sensors are used
to measure atmospheres in confined spaces. Chris and his team
have access to newer techniques. In addition to the
microdrone mentioned above they can use infra-red cameras to
identify gas clouds.
Chris's team and other HSL scientists provide a vital role in
gathering samples and data at the scene of an incident.
Underpinning this front line work are the resources back at
These include the Analytical Sciences Unit. Scientists
such as Dr Duncan Rimmer (head of the unit) and team leader Ian
Pengelly painstakingly work to characterise samples retrieved from
site of the incident.
Sources of samples analysed by the unit read like a who's who of
high profile UK health and safety incidents. They include
Buncefield, ICL Plastics, Ladbroke Grove, , the King's Cross fire
and the Marks and Spencer asbestos exposure case.
This work requires state of the art analytical equipment.
The workhorse is gas chromatography (which separate substances in a
mixture) coupled to mass spectrometry (which identifies them).
GC-MS is highly sensitive and quantitative. Automated thermal
desorption (ATD) is often used for volatile organic compounds. A
relatively new feature of ATD is the ability to split off and
retain a portion of the sample during the analysis for later
"We retain samples for as long as required, which may be five
years or more," says Dr Duncan Rimmer. "They can then be made
available for verification should, for example, our original
analytical result be challenged prior to court proceedings."
Examples of samples analysed by the unit include:
- Pieces of machinery from a fairground accident
analysed for the presence, or in this case the absence of lubricant
(maintenance failure is a cause of many fairground rise
- Plastic from failed guarding on machinery (to check
that cheap plastic has not been substituted for the tough
polycarbonate which should have been used).
The results of the analyses can sometimes be surprising.
In one case a tank containing formic acid leaked through a tap
fitted to the tank. Analysis confirmed that the tank and tap
were made from polypropylene, the right material as it is resistant
to formic acid. However, on dismantling the tap HSL analysts
discovered that the washer was apparently missing. At first
this seemed an obvious cause of the leak. However, sharp
analytical eyes noticed a thin smear of pale residue on the
tap. Further analytical work established that this was the
remains of a washer made not of polypropylene but nylon.
Normally this should not have mattered because nylon is very
resistant to most common organic compounds. Unfortunately,
formic acid is one of the exceptions. Over time it had slowly
dissolved the nylon washer. This example illustrates the
importance of meticulous attention to detail and looking beyond the
Big bang theories
Explosions and fires feature among the more spectacular and high
profile disasters investigated by HSL. John Hodges is a team
leader in the Explosives Atmospheres Section. Elucidation of
the cause of an explosion can often involve a certain amount of
"In 2004 an explosion and fire in the ICL plastics factory
(commonly referred to as the Stockline plastics factory explosion)
near Glasgow killed nine people. It took over 300 fire and
rescue personnel to extinguish the blaze."
"When we arrived we had no idea what had led to the
collapse. A building had collapsed. But was this the
result or cause of the explosion? And was the explosion itself a
plastic dust or gas explosion?"
A possible source of gas was an underground pipeline leading
from a liquid petroleum gas (LPG) tank.
"We forced smoke with pressurised air into the underground
pipeline. This revealed leaks in the pipework. One particular
leak occurred at the point where the pipe went into the
building. What had happened over the years was that the
ground level had been gradually built up over the pipe. A
concrete slab was placed over where the pipe was to provide a
pathway for fork lift trucks. Corrosion, together with the
resulting pressure caused cracks in the pipe. Our smoke test
demonstrated that escaping gas was drawn into the basement through
the hole in the wall where the gas pipe itself entered the
It was maintenance failures that caused the ICL plastics
fire. But sometimes the risks can be increased by
"improvements". For example, over the past twenty years or so
intermediate bulk containers (IBCs) have steadily replaced steel
drums for transporting liquids. The advantages of these cubic
plastic containers include resistance to corrosion, ease of storage
and ease of recycling.
Unfortunately a series of incidents in the UK indicated very
poor performance when exposed to fire. In some cases it was
possible to ignite a container with a single match.
"This was surprising," says Dr Graham Atkinson of HSE's Fire
Engineering Section. "Tests carried out in the USA
indicated good performance when the containers were exposed to
fire. However, when we investigated we discovered that these
tests had been carried on containers filled with water. When
filled with other liquids fire performance significantly
Research by Graham Atkinson and his team confirmed this
effect. They not only showed that resistance to fire was
dependent on the nature of the liquids contained; they were
able to show which liquids were most likely to cause problems.
These included hydrocarbons such as fuel oils, edible oils
and lubricants. In 2006 Dr Atkinson co-authored a
paper1 on this work which was awarded the Frank Lees
Unravelling the cause of an accident is useful for apportioning
responsibility. More importantly it also provides
understanding which can help prevent a recurrence.
An example is the recent work prompted by the 2005 Buncefield
incident. Buncefield occurred when vapour from an overflowing
fuel tank ignited. Graham Atkinson applied the science of
fluid dynamics (which describes fluids, e.g. gases, liquids and
vapours in motion) to see what lessons could be learned.
An overflowing fuel tank results in a cascade of fuel droplets
falling through the air. The droplets drive a flow of
fuel-contaminated air downwards and the result is a heavy,
explosible vapour cloud that rolls away from the tank.
Using fluid dynamics Graham Atkinson and his team
demonstrated how such a heavy vapour cloud would move through a
fuel storage depot. They showed how factors such as the
nature and height of the tank and the rate and volume of overfill
combine to determine the rate and extent of the vapour cloud's
movement. This information now helps the HSE advice local
councils on planning applications. The councils can now
more confidently specify how close property developments can be
allowed to approach fuel storage depots.
Dr Mat Ivings, HSL's head of Computational Fluid Dynamics (CFD)
section aims to take fluid dynamics one step further.
CFD draws on vast amounts of computer processing power to solve the
partial differential equations governing fluid flow. The maths is
incredibly complex and not surprisingly, all seven members of Mat's
team have PhDs in fluid dynamics. But the end result is an
animated and highly visual picture of the flow.
"We can use this to understand which factors control the flow
In 2007, Mat and his team used CFD to understand how leaks from
the Pirbright Institute of Animal Health facility carried foot and
mouth disease (FMD) virus into the surrounding countryside leading
to the 2007 FMD outbreak.
"The drains had been inspected and found to be in poor
condition. In places tree roots had penetrated them. It
was likely that the spread of the virus originated from a leak of
effluent from a manhole. The source of the contamination was
a laboratory within the facility which handled the FMD virus.
We ran our model of the drainage system for different effluent
pumping rates and were able to demonstrate that the overflow
occurred with the higher flow rate."
CFD has also been used to model smoke movement from a fire in a
London underground station and predict the dispersion pattern of a
gas release in an enclosure.
And more recently it has thrown further light on
Buncefield. CFD modelling showed that obstacles such as
hedges and buildings, and the slope of the ground have
significant effects on the movement of heavy flammable vapour
clouds. To assess the consequences of a Buncefield-type tank
overfilling release at other fuel depots, the CFD team have been
running simulations, using a complex three-dimensional model that
incorporates terrain data supplied by the Ordnance Survey. Their
results have provided valuable information to help design
mitigation measures and reduce the risks arising from such
Virtual modelling initiatives such as CFD are providing valuable
information already and promise even more for the future. But
HSL still make use of hard physical evidence. Examples
include the use of tower cranes to investigate collapses and study
the effect of wind on crane components (Richard Isherwood of HSL's
Engineering and Personal Safety Section) and sections of rail to
determine train accidents (Bill Geary of the metallurgy).
Bill Geary points to a section of rail retrieved from the
Hatfield train derailment. "The cracks visible in the
top of the rail are caused by the localised very high stresses
between the train wheel and the rail. Initially, the cracks grow at
a shallow angle but the cracks continue downwards into the rail
until a critical size is reached, at which point catastrophic
failure of the rail occurs. This is what derailed the train at
Over time harder rails have been used to reduce the wear rate
and thus costs, however, this has led to an increase in incidences
of rolling contact fatigue cracking. Since Hatfield sophisticated
inspection techniques have been developed to detect rail head
cracking and rail grinding is used to remove cracks before they
reach a critical size."
HSL have a sustained track (no pun intended) record in accident
investigation. What is their secret? Obviously the quality of
staff is an important factor. A great number of them are
recognised as world experts in their fields.
But equally important is the synergy resulting from the presence
of all these scientists on one 550 acre site.
They are able to bring a multi-disciplined approach to
accident investigation. These range from "hard" sciences such
as chemistry, mechanical engineering, physics and mathematics to
the so called "soft" sciences such as psychology and the social
sciences. This means that when HSE and other organisations
engage HSL they are not just buying into a single expertise.
They are hiring science.
1 CONTROLLING THE FIRE RISKS
FROM COMPOSITE IBCS, G. Atkinson and N. Riley, Symposium Series