Professional Documents
Culture Documents
LPB 297 Online
LPB 297 Online
Flixborough
Issue 297, June 2024
DS34HAZARDS34HAZARDS34HAZA E
Safet
Hazards34
ISC
re
y
t
n
Ce
DS34HAZARDS34HAZARDS34HAZA
DS34HAZARDS34HAZARDS34HAZA
DS34HAZARDS34HAZARDS34HAZA
DS34HAZARDS34HAZARDS34HAZA
5–7 November 2024, Manchester, UK
DS34HAZARDS34HAZARDS34HAZA
Join the major hazards community at Hazards 34 to review good practice, current thinking, lessons learned
and emerging challenges in process safety and hazard management.
DS34HAZARDS34HAZARDS34HAZA
Hazards 34 is an industry-focused event ideal for anyone who is active in process safety and risk management
for chemical process facilities or other facilities dealing with hazardous materials. It is an exciting chance to
DS34HAZARDS34HAZARDS34HAZA
connect with fellow professionals, build networks and share insight and experiences in person.
Programme themes
DS34HAZARDS34HAZARDS34HAZA
Q Assurance
Competencies
Q
Q
Hazardous Area Classification
Human Factors
Q
Q
New Energies
Non-routine Operations
DS34HAZARDS34HAZARDS34HAZA
Q
DS34HAZARDS34HAZARDS34HAZA
Q Cross-sector Learning (Water) Q Hydrogen Lessons Learned Q Safety Culture - Leadership
Q Digitalisation Q Hydrogen Risk Assessment Q Safety Culture - Operator’s
Learning from Incidents Experience
DS34HAZARDS34HAZARDS34HAZA
Q Engineering & Design Q
Speakers
DS34HAZARDS34HAZARDS34HAZA Dr Marlene Kanga Fiona Macleod Andrew Curran
DS34HAZARDS34HAZARDS34HAZA
Rux Energy Independent Health & Safety Executive
Trevor Kletz Hazards Consultant New Technologies,
Lecture Bhopal, 40 Years On Developments and
DS34HAZARDS34HAZARDS34HAZA
The Challenges Energy Transition
Facing the Process
Industries
DS34HAZARDS34HAZARDS34HAZA
Natech Events, the H2: the Risks and the
Risks to Process Opportunities
Safety and the
DS34HAZARDS34HAZARDS34HAZA
Environment
DS34HAZARDS34HAZARDS34HAZA
www.icheme.org/hazards34
Find out more and register:
DS34HAZARDS34HAZARDS34HAZA
Gold Sponsor Silver Sponsors
DS34HAZARDS34HAZARDS34HAZA
LPB 292
Contents
2 Editorial 17 Building-in in-building
Loss Prevention Bulletin Fiona Macleod, Chair of the safety
Loss Prevention Bulletin Editorial Andrea Longley and Ken Patterson
Articles and case studies Board, introduces this special outline how the siting and design
from around the world issue commemorating the 50th of control rooms and other
anniversary of the Flixborough occupied buildings were factors in
Issue 297, June 2024 disaster. Flixborough and several later major
incidents, and how this led to the
Editor: Tracey Donaldson 3 Flixborough – in the development of the UK Chemical
Publications Director: words of the witnesses Industries Association’s Guidance
Claudia Flavell-While Rick Loudon, author of Flixborough on Occupied Buildings.
Subscriptions: Hannah Rourke – I was there, recounts the events
Designer: Alex Revell
of 01 June 1974 and its immediate 22 The effect of control room
aftermath, using the words of location, architectural
Copyright: The Institution of Chemical the emergency services and local
Engineers 2024. A Registered Charity in population who witnessed the
design, systems
England and Wales and a charity registered disaster. configuration and human
in Scotland (SCO39661) factors on major accident
ISSN 0260-9576/24
7 Nylon Years – consequence and
An interview with likelihood
The information included in lpb is given in
good faith but without any liability on the
Ramin Abhari Andy Brazier discusses the
part of IChemE Ramin Abhari talks about what importance of overall design of
inspired him to write his graphic the control room and associated
Photocopying novel, Nylon Years – and what systems highlighting that more
lpb and the individual articles are protected are the key lessons that he would focus is needed on supporting the
by copyright. Users are permitted to like readers to take away from the control room operator to perform
make single photocopies of single articles story. their critical role of keeping the
for personal use as allowed by national
plant safe and responding to
copyright laws. For all other photocopying
permission must be obtained and a fee 11 Flixborough — accidents.
paid. Permissions may be sought directly remembering ICI’s
from the Institution of Chemical Engineers,
response to the disaster 25 Recollections of
or users may clear permissions and make
Frank Crawley remembers, as Flixborough
payments through their local Reproduction
a young production manager Recollections from IChemE
Rights Organisation. In the UK apply
to the Copyright Licensing agency in 1974, how the Flixborough members of Flixborough and how
Rapid Clearance Service (CLARCS), 90 disaster caused ICI to undertake the disaster ultimately shaped the
Tottenham Court Road, London, W1P a review - led by Trevor Kletz - of progression of their careers.
0LP (Phone: 020 7631 5500). In the USA its own processes and implement
apply to the Copyright Clearance Center new process safety improvement 33 Safer leader page:
(CCC), 222 Rosewood Drive, Danvers, MA
01923 (Phone: (978) 7508400, Fax: (978)
strategies. An interview with
7504744). Ken Rivers
13 Flixborough and inherent Ken Rivers shares his perspective
Multiple copying of the contents of safety – inspired by on process safety and the
this publication without permission is
Trevor Kletz importance of good leadership,
always illegal.
Andy Brazier explains how the drawing on learnings from
Institution of Chemical Engineers concept of inherent safety – for Buncefield, Flixborough and other
Davis Building, Railway Terrace, example, what you don’t have, accidents.
Rugby, Warks, CV21 3HQ, UK
can’t leak – came to the forefront
Tel: +44 (0) 1788 578214 in process design considerations
Fax: +44 (0) 1788 560833 following the disaster.
Email: tdonaldson@icheme.org
or journals@icheme.org
www.icheme.org
Editorial
Flixborough – 50 years on
Fiona Macleod
The June 2024 edition of Loss Prevention Bulletin marks the and Ken Patterson and Andrea Longley draw attention to the
50th anniversary of the Flixborough accident. importance of building-in in-building safety. There is a delicate
On 01 June 1974, a huge cloud of cyclohexane exploded balance between having the right people in the right place to
above the Nypro factory, causing 28 deaths, multiple injuries prevent accidents, while minimising the occupancy in potential
and widespread damage. danger zones.
Rick Loudon, a member of the fire service, gives a unique IChemE members recount where they were on the afternoon
insight into the emergency response in his paper Flixborough of Saturday 01 June 1974, and how the accident changed their
- in the words of the witnesses. First responders were met professional lives. Finally, in our safer leader page, Ken Rivers
with scenes of utter devastation: villages destroyed, fallen shares his perspective on the importance of good leadership
high voltage cables arcing on the ground, a wall of flame in process safety, drawing on learnings from Buncefield,
and a growing risk of escalation -— the British Steel blast Flixborough and other accidents.
furnaces north of Scunthorpe were in danger of overheating The tragic accident at Flixborough catalysed the passing of
after a critical cooling water main was shattered by the Nypro the Health and Safety at Work Act and started the international
explosion. journey to the Safety Case approach to process safety in high
The court of inquiry found that the root cause of the hazard industries.
accident was an ill-judged modification to the cyclohexane There are no new accidents. Unless we make the effort to
plant, compromising the mechanical integrity and leading to a understand what happened
catastrophic loss of containment. and why, we are destined
Although crucially important in high hazard industries, to repeat them. The 28
the lessons from Flixborough shouldn’t be confined to people who died and
Management of Change. Ramin Abhari discusses the all those physically and
inspiration for his graphic novel Nylon Years which illustrates mentally injured, paid far
the influence of political, economic, techological and social too high a price for us to
change, highlights the importance of listening to the safety ignore the lessons from
concerns of workers, and reminds us of the importance of Flixborough.
inherent safety in design.
Frank Crawley remembers the ICI response to the accident Fiona Macleod
and Andy Brazier expands on Trevor Kletz’s approach to Chair, LPB
inherent safety. Andy also writes about control room design Editorial Panel
Incident
knowledge and
to check if the other five reactors had a similar fault’) and (© Andrew Henry)
competence
bypass it to keep production going. The following is taken from
the official enquiry and is not supposition: tonnes of Cyclohexane gas were released to the atmosphere,
very quickly finding a nearby ignition source and causing a
This bypass was constructed on site using material
Vapour Cloud Explosion (VCE) that at 17.05 registered at a
available, and connecting bellows for each end were
height of 200 miles and caused an RAF Canberra Jet flying at
engineering
and design
supplied from a specialist company. Whilst the fabricators
6000 feet to alter course.
who constructed the bypass did exactly as they were
The explosion caused widespread damage in the
instructed, the bypass was completely untested and failed
surrounding area and plate glass shop windows in Scunthorpe
to comply with the relevant British Standards at the time
five miles away were blown in causing widespread damage
and did not meet the bellows manufacturers specific
and injuries to shoppers from flying glass. In the village of
systems and
installation instructions.
procedures
Flixborough itself, out of 79 properties only seven did not
It should also be noted that the official enquiry into the incident suffer damage.
stated: Fire appliances were mobilised from all over the area with
the furthest away on the initial attendance that day coming
The key post of ‘Works Maintenance Engineer’ was vacant
from as far away as Hyde in Greater Manchester and Ossett
and had been since early 1974.
and Wakefield in West Yorkshire (bearing in mind there were
It also be noted that: ‘None of the engineers on the workforce no motorways in the area in 1974).
had an engineering background, all were chemical engineers The explosion had also destroyed a cooling main from the
and were incapable (through qualification) of recognising River Trent to British Steels Normanby Park Steelworks and the
what was, in essence ‘a simple engineering problem’ let alone blast furnaces needed thousands of gallons of water supplying
human factors
resolving it’. (Please be aware these statements were made to prevent an even bigger explosion occurring which would
by the team at the official enquiry into the incident nearly fifty have destroyed a good proportion of northern Scunthorpe!
years ago. They are not intended to be a criticism of current Ambulances and Police came from a wide area and London
practice, but cannot be altered from the official report of the Fire Brigade and London Ambulance Service offered assistance
time.) which was not ultimately needed.
Fortunately, as it was a weekend, many staff from the offices
A member of the workforce said:
etc. were at home and contractors working on the site had
finished work by 12 noon. ‘I saw the flames shooting about 250 feet in the air from
At 1653 hours the bypass failed and approximately 16 section 25A, then I heard a siren sounding, I put on my
Figure 2: General view from the banks of the River Trent some Figure 3: Fire-fighters at no.5 reactors (© Scunthorpe
distance away (© Andrew Henry) Telegraph)
safety helmet and went to the door but then there was slippers and with no crash helmet on as he’d rushed out of
another huge bang no more than 30 seconds after the first the house to find us and to see what had happened!’
bang. All I could see was a big sheet of yellow flame in the Another young girl was in her parents’ car coming home from
direction of section 25 and then I was thrown onto the floor holiday and recalled:
by the blast. I could hear burning and some other bangs
and I tried to crawl in the opposite direction. I was shouting ‘As we got nearer to the River Trent we suddenly saw
for help as I crawled and I finally came to a fence and could a huge blue flash shoot into the air to the north, in the
go no further. I was very tired and decided to lay where I vicinity of the Nypro works followed by a huge explosion.
was and go to sleep, then somebody came to me.’ We didn’t know what it was or if there was a danger of
toxic fumes. Dad told us to shut the car windows for the
Another employee who was working overtime that day said:
rest of our journey home!’
‘I was working overtime and as was normal practice, if we
worked the full shift we got paid for that shift; however, The first crews on the scene (four appliances and 12 firemen)
if we finished our job early, we had the choice of either were met with a scene of utter devastation and the first
staying for the remainder of the shift and getting paid or message back was:
‘knocking off early’ and claiming time in lieu. I needed to ‘Major Disaster, approximately 40 acres of a chemical
go into town, so I took the time in lieu option and left work works completely involved in fire, many casualties and an
around 16.10 hours and went home for a bath before unknown number of persons unaccounted for.’
heading into town. I’d just got in the bath when there was
an almighty bang! My first thought was Mum’s blown the
cooker up in the kitchen! I found mum outside talking to
a neighbour about what had happened (we lived on the
northern edge of Scunthorpe at the time and had a clear
view across the fields of what was occurring about 4 miles
away). Later it transpired that whilst mum hadn’t blown
the cooker up, our house and many others in the area had
suffered significant structural damage and we required a
new roof. So, my decision to claim a ‘lieu day’ as opposed
to cash that day was quite a significant moment in my life!’
Interview
knowledge and
refused, and the miners started a “no overtime strike.” When the
competence
company, the plant that our team built based on a new process we
Arab oil embargo took effect later that year, the Heath government
had developed experienced two fire incidents involving injuries.
instituted measures requiring the largest industrial energy users
That’s when I decided to write something about process safety.
to cut back their power consumption. This led to the decision by
Since I had read that stories and images are remembered longer,
Nypro to turn off the electric powered reactor agitators.
I chose to tell the story of a chemical plant disaster in the graphic
engineering
novel format. I chose Flixborough as the inspiration for the story.
and design
Can you explain what is meant by ‘The Water
What did you base your research on? Hypothesis’?
The Court of Inquiry report was my main source for the factual When two immiscible phases are heated in the same tank, each
events surrounding the accident. For technical information, I exerts its own vapour pressure. This phenomenon is encountered
when a hot hydrocarbon is transferred to a tank that contains
systems and
procedures
relied on King’s writings about heat up of the two-phase water-
cyclohexane system, but also on other Flixborough-inspired papers hydrocarbon with a pool of water at the bottom. When the settled
on nitrate stress corrosion cracking and the hydrodynamics of the water layer rapidly heats up as a result, a sudden eruption of the
“bypass pipe” failure. Other research included original patents liquid can result. To avoid this, the rule of thumb is to ensure the
filed by Stamicarbon/DSM on cyclohexane oxidation technology tank pressure is greater than the sum of the vapour pressures of
and chemical market reports about caprolactam and nylon-6 in the the hydrocarbon and water.
‘60s-70s. Since I was writing about 1970s northern England, I also During hot circulation in the Nypro CSTRs, the hot cyclohexane
researched the time, culture, and politics of that time and place. flowing over the water layer in R-4 would have heated up the latter
in a similar manner. According to King’s analysis, after the R-4
Did you agree with the findings of the (1975) cyclohexane reached 155⁰C, the water layer would have needed
Court of Inquiry on the root cause? to reach 145⁰C for the sum of the vapour pressures to exceed the
operating pressure of 7.8 bar. When that condition was met, a
The Court of Inquiry report is an extensive accident investigation
pressure spike with eruption of liquid could have exerted stress on
based on an exhaustive number of interviews and data collection.
the bypass pipe assembly, initiating the squirming of the pipe at
I don’t think anyone disagrees with the “bypass pipe” as being
the bellows and the rupture that followed.
the root cause. However, I agree with King that the failure of
the “bypass pipe” could have been initiated by the water pool
Do we know for sure that the concentration of
culture
Phenol
Cyclohexanone Nylon-6
Caprolactam
New Nypro
Process Unit
(1974)
Cyclohexane
To Flare
Air
Nitrogen
(1)
(1)
(1) (1) (1)
Scrubber/Absorber
Distillation Train
Cyclohexane Recycle
Water
Cyclohexanone to
Caprolactam Unit
Cyclohexane
Air
Nitrogen
Figure 3 – The rector system at the time of the accident, showing how a pool of water
byproduct could form in unagitated R-4 reactor next to the “bypass pipe” assembly
reactor behaviour. Since cyclohexane oxidation forms water, its joint venture in England, they also built an identical caprolactam
concentration in the reactor system increases with conversion. plant in Augusta, Georgia (USA). In 2016, when the Augusta
For all first order or higher reactions, the conversion curve caprolactam plant was permanently shut down, I used Google
(conversion as a function of reactor number for CSTRs in series) Earth to zoom inside that plant. I could spot three cyclohexane
is expected to look something like what the operators have oxidation trains identical to what was in operation at Flixborough
sketched in the bottom panel of Page 20 of Nylon Years. That is, (same six CSTRs in series with each reactor at a higher elevation
the concentration of water would be lowest in R-1 and highest than the next). Clearly, the technology continued to be used for
in R-6. However, since most reaction happens in the first 4-5 caprolactam production and capacity expansion.
reactors, the increase from R-4 to R-6 would have been minor.
(That’s why when R-5 was pulled out, there was no significant What lessons would you like us to take away from
drop in conversion performance.) Furthermore, since the last your story?
reactors would have been the hottest (highest conversion of the
exothermic oxidation reaction without heat removal), a case can Lesson 1. Inherent safety in process design.
be made that some of the water may have been stripped with the
sparging air such that during normal steady-state operation, R-4 Due to low per-pass conversion, large volumes of cyclohexane
would have had the highest water content when the agitation was had to be circulated at elevated temperatures and pressure.
stopped.
Lesson 2. Weak signals and normalisation of
None of the shift team or control room records deviation
survived the accident — how much of the Although product quality and production rates were not
description of operator actions leading up to the impacted, running with none of the reactors stirred seems to have
explosion in Chapter 10 can you be sure of?
That part is complete dramatisation. There is no doubt that
the operators were immediately made aware of the loss of
containment as reactors rapidly depressurised. What they did in
the two minutes between that, and the vapour cloud explosion is
something we can only imagine. I read some firsthand account by
a technician working in the lab that he saw a haze (of cyclohexane
vapour) as he was evacuating the plant with his colleagues, before
he was knocked down by the vapour cloud explosion. I imagined
something similar with the process operators, except that in my
story they chose to mitigate the vapour release as they had done
with much smaller vapour releases in the past.
Incident
engineering
and design
What was the outcome? A number of changes were
As it was a Saturday I was at home with my wife and young
required. Table 1 describes the changes, the reason and the
family when news of the Flixborough explosion hit the evening
actual requirements. This is a sample of the more significant
headlines. The footage of incident was initially unclear, but
changes.
the papers next morning were more informative and from
The first change was quite simple. The emergency
an ICI methodology, written by Bill High (which became part
systems and
procedures
procedures did not cater for an event of the size of
of the IChemE Monograph on explosions) and placing the
Flixborough.
epicentre on the Cyclohexane Oxidation Plant, it was possible
to assess a likely TNT equivalence from the damage profiles The second change was self-evident from the number
seen in the papers. It was quite unbelievable and at the best of joint leakages and the joint damage. It only required
estimate it appeared that the leak rate to cause that explosion specification of an acid resistant and solvent tolerant jointing
was the equivalent of a manhole cover being blown off. One material.
of the anomalies was the off-site damage to houses. This was The third change was installation of hydrocarbon gas
put down to a second explosion (and a new epicentre) in the detectors which were then coming onto the market and had
Caprolactam Plant been used successfully elsewhere in ICI.
As Monday 03 June was a bank holiday I decided not to go The fourth change was based on the observation that most
to the works as it would have inferred concerns. fire-water drainage systems were inadequate. Fire zones were
Things happened quickly. First came a request from Trevor defined and the probable fire-water loading was assessed for
Kletz to review the improvements that I felt were necessary each zone. Then water was pumped into each zone and “back
on the Cyclohexane Oxidation Plant (KA, ketone alcohol, at up” and segregation was assessed. The results were worse
Ardeer). This was already a part of my role. The second was than anticipated. Pinch points were treated by bunding and
a presentation by D Waters, the Fire Chief at ICI Wilton (who fire-water spread by grading. Fire water recovery required the
I knew well from my production “fire experiences”). He was construction of an earth settling area and strategically placed
seconded to organise the firefighting at Flixborough but he pump-out points.
produced a photo of a piece of piping, bent like a drinking Strangely there were two improvements which came from
Review all safety procedures and Most procedures were based on upsets A whole raft of new emergency procedures
upgrade in the light of Flixborough and not emergencies was drawn up
Change the piping jointing Cyclohexane dissolved the elastomer in Replace all joints with SEW loaded with
Compressed Asbestos Fibre joints which “graphite” joints. The joint comprised an
then leak inner and outer retaining ring filled with a
spiral of graphine and stainless steel sealed
to the rings.
Upgrade the Fire and Gas detection The installed system did not include gas A new gas detector system, strategically
system detectors which were just coming on the placed, was installed
market
From earlier observations it was clear A simulated full fire water load test was The concrete was regraded, mini buds were
that the fire-water drainage was not fit carried out to simulate a full fire. installed to limit fire spread on water and
for purpose in most chemical plants pump-out points identified with “settling
tank or bunded areas”
The process used a slurry of boric acid Installed secondary flexible pump seals CRANE (the pump seal manufacturer) were
which damaged pump seals between the flush point and the impellor consulted and were just bringing to the
market “tetra-lip seals These were installed
The installed Emergency Isolation The Oxidation process results in a strong New valves were purchased and installed
Valves were not reliable acidic environment but the seals must
also be acid resistant and tolerant of the
boric acid slurry
writing a specification for each case/item but not knowing if it convinced that they could supply and added an extra feature
knowledge and
competence
existed. – a scrapper ring to dislodge any slurry off the ball surface.
The fifth change was again the result of pump seal leakage. This was the first ever use of their valves in Europe but now
Seal flushes were not perfect as boric acid could still migrate to NELES supply nearly all sub-sea pipeline isolation valves in the
the seal face. A specification was written for a secondary seal, offshore O & G industry.
between the impellor and the main pump seal, using a flexible There were other minor changes and one failure.
engineering
and design
graphite loaded PTFE seal which was to be an annular fixture Emergency depressing could not be done properly as there
lubricated by the seal flush but roughly shaped as a comma was no flare stack on the plant. Dispersions calculations
in cross section such that it would give some sealing should showed that venting to atmosphere was viable (it would not
the main seal collapse. CRANE (the pump seal manufacturer) now be so due to the environmental considerations) but the
were consulted and given a specification of what we wanted. integrity could not be proven so it was abandoned.
CRANE were a little surprised as they were just about to bring
systems and
Did it make a difference? Yes. Operation was easier,
procedures
their new secondary seal to the market and showed us a emergency procedures were sharpened up, leak frequency
prototype of their TETRA-LIP SEAL at the meeting which was more than halved — in particular on pumps and pipe jointing,
exactly as we had specified!! emergency isolation was far more reliable and,possibly of
The final change was based on wish list for the “specification more importance, the operators felt more confident in the
of all that we wanted in a ball valve”. A valve purchaser took safety of the plant.
on the task and identified a Finnish company by the name History is such that about ten years later the Nylon business
NELES which could manufacture to that specification. was sold to Dupont and eventually the whole business was
After a trip to Finland in a very cold February I was discontinued ten years later.
Safety practice
knowledge and
competence
in the hierarchy of risk controls. This paper is a shortened safety he proposed the following very simple but effective
version. statements5:
Keywords: Inherent safety, Flixborough, Kletz • “what you don’t have, can’t leak”
• “people who are not there can’t be killed”
engineering
and design
What is inherent safety? • “the more complicated a system becomes, the more
It is probably fair to say that inherent safety is a concept rather opportunities there are for equipment failure and human
than a clearly defined method or approach. This may explain error”
why the development of a universally agreed definition has not The best way of preventing a leak of hazardous material is to
been straightforward. use so little that it does not matter if it all leaks out, or to use a
systems and
procedures
Whilst Kletz wrote a lot about the subject, he does not appear safer material instead. We cannot always find ways of doing this
to have used a specific definition. One of the closest attempts but once we start looking for them, we find a surprisingly large
appears in his autobiography3 where he says the main concept number.
is that “it is better to remove a hazard than to keep it under Whilst hazard elimination will always be the most effective
control.” measure, Kletz was very well aware that this was not always
Organisations including US Center for Chemical Process possible or desirable. With this in mind, keeping people away
Safety (CCPS), UK Health and Safety Executive (HSE) and the from hazardous areas can be very effective at reducing the
Energy Institute have actively explored the subject with the consequences of accidents that occur.
following common themes: There is a view that complication is inevitable today.
• risk reduction is an intrinsic part of the process and not an Sometimes it may be, but not always. There are many ways in
added layer; which plants have been made simpler, and thus cheaper and
safer. As with the reduction of stocks, the constraints are often
• it is permanent and inseparable from the process;
procedural rather than technical. We cannot simplify a design if
• it should be balanced with other decision-making criteria, we wait until it is far advanced; we have to consider alternatives
especially where there is significant cost or technical risk. in a structured and systematic way during the early stages of
design.
Relevance to Flixborough With regards to cost, Kletz was adamant that an inherently
Kletz wrote4 “Flixborough in 1974 occurred in a plant for safer plant is also cheaper to build, operate and maintain
the oxidation of cyclohexane with air, at about 150°C and because it can be smaller and use less protective equipment.
Intensification or minimisation
The aim here is to perform the same activity with smaller quantities of hazardous
material or performing an activity less often. This can be achieved by selecting
different equipment and processes that are more efficient or require smaller hazardous
inventories. Switching from batch to flow reactors can significantly reduce inventories.
Substitution
The aim here is to reduce the hazard severity by replacing a hazardous substance or
a processing route with a less hazardous alternative. Another option is to replace a
procedure with one that presents a lesser hazard. Using safer solvents or choosing
processes that require less hazardous conditions.
Attenuation or moderation
The aim here is to use a substance in a way that reduces its hazardous properties or to
use less severe processing conditions. Another way is to store or transport material in a
less hazardous form. It can be achieved by controlling operating temperature to below
where a runaway reaction can occur and storing materials in less hazardous forms (e.g.
paste instead of powder).
Simplification
The aim here is to reduce the likelihood of an accident through inherent features of the
design. This can involve designing processes, equipment and procedures to eliminate
opportunities for failure, including human error; also, designing equipment that cannot
be exposed to extreme process conditions by the worst-case processing conditions.
Table 1: A series of images Kletz presented to illustrate the principles of inherent safety6.
Principles of inherent safety these changes earlier is likely to be cheaper and cause fewer
knock-on issues.
In his workshop notes published by IChemE in 1978, Kletz
references Edward de Bono as saying simple pictures can be In the very early stages of a project decisions are made
very powerful at conveying ideas. Images do not have to be about what to make, by what route and where the facility
accurate or descriptive, but simple enough to lodge in the will be located. Adopting and mandating formal conceptual
memory. Above is a series of images Kletz presented to illustrate stage studies can ensure sensible discussions take place so
the principles of inherent safety6. that optimal decisions can be made. Researching all available
chemical processes, including low-inventory flow reactions and
Elimination may be considered the most fundamental
semi-batch methods, and conducting laboratory and pilot plant
principle of inherent safety but did not appear on Kletz’s list
experiments should be considered to ensure the safest chemical
because he generally saw it as a result of applying inherent
process is selected. Also, it sets the scene for the remainder of
safety rather than a principle in itself2.
the project.
During Front End Engineering Design (FEED) or Define
Applying inherent safety through design
phase, when a flowsheet that identifies the main sub-systems
The concept and principles of inherent safety can be applied has been developed, the following can be used as a prompt7:
at all stages of a system’s lifecycle. However, the greatest
opportunities for risk reduction are found at the earlier stages • materials — develop an inventory, identify their hazards and
of development because there are more options to eliminate or consider options to remove or reduce;
significantly reduce hazards by changing the chemical process, • reaction — size of reactors and opportunities to reduce;
fundamental engineering design or plant location. Also, making process conditions and opportunities to make less severe;
systems where it is the duty holder’s responsibility to for convenience rather than necessity (what you don’t have
demonstrate that they are managing their risks rather than can’t leak). At BP Texas City, most of the 15 people who died
prescriptive ‘rule-setting’ systems where the regulator has were in a temporary building that could have been located in
a greater role in saying how risks shall be managed. Other a far safer place on the site (people who are not there can’t be
countries including the US have avoided adoption of the ALARP killed). At Esso Longford the heat exchanger that failed had not
principle. One of the reasons is that it is difficult to define what is been designed to withstand the low temperatures possible under
considered as reasonably practicable for a given circumstance. abnormal or fault conditions (the more complicated a system
The UK’s HSE provides guidance on how to apply ALARP in becomes, the more opportunities there are for equipment failure
practice. Cost benefit analysis may be one approach but can and human error).
be complicated and relies on quantified data that may not be Controlling risk is not simple. Opportunities for reduction
readily available in any useful form2. should always be looked for, whilst being aware of unintended
Guidance for permissioning within the Control of Major consequences. Whilst it may be easier at the early stages of a
Accident Hazards (COMAH) regulations states that “ALARP design project, the principles of inherent safety can be applied at
demonstration for individual risks is essentially a simple concept any time. When contemplating a task everyone involved should
which can be satisfied by the operator answering the following be asking themselves whether all reasonably practicable steps
fundamental questions”9. have been taken to remove hazards, if people who do not need
to be present have been kept away and if arrangements are as
1. What more can I do to reduce the risks?
simple as they could be.
2. Why have I not done it?
An inherently safer solution may not actually create the lowest
Answers to the first question are qualitative in nature and overall risk. Applying the hierarchy of risk controls is not a case
involve looking systematically at the risks and drawing up, of selecting which control to apply but can provide a structured
in a proportionate way, a list of measures which could be way of evaluating the potential strengths and weaknesses of
implemented to reduce those risks. different options. Ultimately the aim is to achieve risks that are
The answer to the second question may be qualitative or ALARP, which requires you to continually consider what more
quantitative in nature depending on the predicted level of risk can be done to reduce risk and demonstrate that doing more is
prior to the implementation of those identified further measures. not beneficial.
The guidance states that if “it cannot be shown that the cost All this is taking place in a global context. We may feel that
of the measure is grossly disproportionate to the benefit to our responsibility is to the safety of our colleagues, neighbours
be gained, then the operator is duty bound to implement that and local environment; and that decisions we make that may
measure”9. However, there are often reasons to not implement affect risk in another part of the world are not our concern. But
additional measures that are not purely due to financial cost. morally we all have to be aware of how the decisions we make
Risk transferral is very often a factor where a measure to reduce affect others. The message for industry is that it should “export
one risk increases another. inherent safety not risk.”10
In some cases, ALARP can mean that an inherently safer
solution is not safer overall. For example, choosing to not References
make a product, to eliminate a hazard, may simply mean that
production is moved to another site, possibly in another country. 1. The Flixborough Disaster. Report of the Court of Inquiry.
The alternative may apply lower safety standards. Also, risks Department of Employment (1975)
of transport will have increased. In this case the issue may be 2. Brazier, A. Edwards, D. Macleod, F. Skinner, C. Vince, I.
moral rather than economic, and there may be an argument to Trevor Kletz Compendium. Elsevier (2021)
say that such global issues are not necessarily the responsibility 3. T. Kletz, By accident … a life preventing them in industry, PFV
of commercial organisations. However, with increased scrutiny Publications (2000).
from customers of the supply chains of their suppliers it is 4. T. Kletz, Plant Design for Safety - a user friendly approach,
possible that keeping production local may be the best solution Hemisphere Publishing Corporation (1991)
from all perspectives.
5. T. Kletz, Lessons from Disaster, Institute of Chemical
Engineers (2003).
Conclusion
6. T. Kletz, Cheaper, Safer Plants or Wealth and Safety at
There have been many publications since the Flixborough Work (Notes on Inherently Safety and Simpler Plants), The
disaster encouraging us to adopt inherent safety, with very Institution of Chemical Engineers (1984)
little (if any) dissent. Similarly, the hierarchy of risk controls is
7. T. Kletz, P. Amyotte, Process Plant, A Handbook for
well established and accepted. However, Safety Instrumented
Inherently Safer Design”, 2nd Edition, Taylor & Francis (2010).
Systems (SIS) have proliferated, which are clearly an add-
on safety device rather than an inherently safe solution. 8. A. Brazier, N. Wise. Making Sure Risks are ALARP. The
Instead of eliminating hazards they can be used to allow more Chemical Engineer (2021)
hazardous processes to take place, whilst also increasing overall 9. Health and Safety Executive HID CI5A. Guidance on ALARP
complexity. Decisions in COMAH. SPC/Permissioning/37. Version
There have been plenty of accidents since Flixborough 3, http://www.hse.gov.uk/foi/internalops/hid_circs/
that would have been avoided or far less serious if inherent permissioning/spc_perm_37/ (Accessed March 2024)
safety had been adopted more widely. At Bhopal the methyl 10. D. Edwards, Export inherent safety - not risk, Loss Prevention
isocyanate that leaked was only an intermediate that was stored Bulletin 240 (2014).
Safety practice
engineering
and design
Mary Evans / The National Archives, London. England.
assurance
Photograph showing the extent of building damage after the Flixborough explosion
process safety and ended elsewhere in safety practice, for As Is Reasonably Practicable” (SFAIRP) — the requirement in the
example in the UK COSHH regulations (dealing with workplace UK’s Health and Safety at Work Act.
chemical exposure). For occupied buildings these can be For those without direct operational roles working in a building
summarised as: Remove, Reduce, Relocate, Protect, and Assess. — which will include control room operators, maintenance staff,
The assessment should be started by obtaining a site layout plant-attached laboratory staff, some technical staff in plant
plan, at a reasonable scale, and marking — colouring in — all management and other supervisory roles, and others — the best
the buildings which have people working in them, together with solution is that the building should be outside the hazard range
the numbers of people in each building. Then the location of the of the potential accidents. This may be possible by relocation of
site’s significant hazards should be marked onto the plan — in a the task or operation to another building, still on-site but outside
different colour. This simple and quite quick process will give a the hazard ranges of the site’s significant hazards. It may also
rapid picture of both the site population and the buildings which be possible to reduce or relocate the source of the hazard, for
are near significant hazards. If a very rough sketch-estimation of example by reducing the on-plant inventory of a flammable or
the likely ranges of the significant hazards is added, the hazard- toxic gas by providing a piped supply from a relatively remote
population interactions will be visible. The hazard ranges will be storage. Any steps of this nature should, of course, be subject
refined and properly calculated as the OBRA process continues to a full management of change assessment to ensure that new
but this simple step will often show both the extent of the problem hazards are not introduced and that overall risks are indeed
and indicate the buildings at greatest risk. It may well suggest reduced. This assessment should take into account the need of
immediate improvements which can be made, and which locations some staff, who may not have direct operational roles, to be near
should be studied first. enough to the plant to ensure good communication and effective
The first task is to remove people from locations and buildings teamwork. It is important that they understand the way the plant
where they are exposed to the inevitable hazards on a chemical works and its actual condition, of both plant and staff. As is often
production site (or other site handling significant quantities of the case in process safety, different risks have to be balanced
hazardous materials). This can often be done relatively quickly and against each other to give the best overall result.
produces immediate gains: at H&W and Texas City a comparison
of who currently worked on the production site and whose work Risk based justification for staff in occupied
actually required proximity to production, showed that around buildings
20% of staff could be moved off site and do their work elsewhere
without any significant detriment. At Texas City this amounted The steps up to this point can be characterised as a hazard-based
to some hundreds of people. Once the types of job have been approach and the CIA guidance advocates that this approach
considered, a second task is asking if it is possible for the numbers should be used first. The aim of the hazard-based approach is to
of people close to plant to be reduced, perhaps by re-defining job identify as many personnel as possible who may be at risk from
roles. Again, overall aggregated human risk can be reduced by potential site incidents and act as quickly as possible to remove
ensuring that only those parts of a job which actually need to be them from locations where they are “in harm’s way”.
done on site are done there, with the other parts carried out in a However, after these steps have been taken, there will usually
safe location. be OBs within the hazard range of potential site incidents which
There will almost certainly still be a significant population are operationally required, and which will need to continue to be
on-site even when the first task is completed. The CIA guidance used. For the buildings in this category, it is necessary to protect
recommends that the next step should be hazard based: the occupants of each OB and assess the risks to ensure that the
considering where people are in relation to the site’s significant residual risk is acceptable — a risk-based approach. The first
hazards. At its simplest, especially for a site which has a Seveso/ of these steps, protect part 1, requires a consideration of all the
COMAH Safety Report, this can be done by firstly calculating hazards for which any building is within the hazard range, and
and then plotting the hazard range for each of site’s major then a demonstration that the building itself does not add to the
hazard events onto the site plan and then highlighting the site risks faced by the people inside it. As examples — is the building
buildings which have people working in them. Figures for the likely to be rapidly affected by a nearby fire and catch fire itself,
outer boundary of the hazard range for various events are given in increasing the risk to those inside? In a foreseeable explosion does
the CIA guidance — for example an overpressure of 30mbar for the building have glass or other materials which could fail and
explosions (pressures below 30mbar do not significantly damage produce dangerous missiles? Is the building’s structure likely to
buildings4); or the distance at which a flammable gas will have fail in such a way as to prevent escape? Are there suitable escape
been diluted to the lower flammable limit concentration. routes from the building to a place of safety? Buildings judged
It is worth noting at this point that those working in direct unsuitable by these tests will need to be either vacated or rapidly
operational roles in a plant are excluded from this part of the OB modified to make them suitable for continued use.
assessment. Indeed, in a site’s overall approach to hazard control, If this first group of tests are passed, the second requirement
the requirement to make the operations of the whole site as safe is to show that, taking into account its location and the identified
as practicable comes before the need for OBRA. If a site operates hazards, the building offers suitable protection to those working
safely and smoothly, everyone’s safety is improved wherever inside it: protect part 2. The assessment needs to take account
they are. The safety of the whole site population is covered by of the consequences to the building if each hazard is realised,
the requirement (on a major hazard site) to demonstrate that “all considering (again as examples) fire resistance, resilience of the
necessary measures” have been taken to avoid major accidents structure to any foreseeable explosion, and means of escape. In
and to ensure that the operational risks have been reduced to “As addition, the protection the building offers to its occupants in the
Low As Reasonably Practicable” (ALARP)11. On other sites this is event of a toxic gas release will need to be assessed, along with
covered by the requirement to operate with risks reduced “So Far any (all the) other identified hazards. These issues (and others) are
Pictures showing buildings rebuilt to withstand calculated maximum credible threats at the Hickson & Welch site. Left, the
rebuilt control room, resistant to a blast overpressure of 500mbar and providing protection in case of toxic gas release (there was
a 40 tonne chlorine storage ~100m away), and right, the site office block, resistant to fire and blast overpressure of 300mbar.
Photographs by Ken Patterson
extremely important in the case of control rooms which are almost which can be used for different hazards and the building design
always required to continue functioning in the event of an accident, techniques which will enable a building to mitigate the various
to enable the plant to be safely shut down. hazards, can be found in the CIA guidance. The use of the
However, the question, as always in process safety is “have CIA guidance and API recommended practice documents 752
we done enough?” The CIA guidance endorses the principle & 753 has been widely discussed in the literature and many
that those working inside an occupied building should not face consultancies offer specialist help in carrying out OBRAs and with
a higher risk than people off-site. The reasoning is clear: these the ALARP demonstration.
are jobs which could be done elsewhere, even though that might
add inconvenience and reduce plant contact. If they could be
done elsewhere, then the buildings being worked in should keep
the people inside as safe as they would be off-site. The HSE
publication Reducing Risks, Protecting People (R2P2)12 gives figure
of 1x10-4 for the risk of death per year as the maximum acceptable
risk level for people off-site and the guidance adopts this figure
for people working in occupied buildings. Of course, this is a
maximum tolerable risk, not a target: the target should be as low as
practicable, preferably 1x10-6 or better.
The description of how to carry out a full OBRA is outside the
scope of this article and is dealt with comprehensively in the CIA
guidance5, though the details will vary from site to site as the
hazards and local conditions very. In the UK an OBRA is expected
as part of a COMAH site’s safety report and forms a significant part
of the site’s ALARP demonstration. However, the CIA guidance
provides the flowchart opposite, which is also included in the one-
page summary5, 6b.
For the ALARP demonstration it is perhaps worth adding two
points:
1. Having calculated the risk level for each hazardous event at
the building boundary, we need to consider the degree of
protection the building provides to give the mitigated risk, that
is the risk the occupants inside the building face taking into
account the protection the building provides;
2. However, it is the sum of the mitigated risks from all the
hazards which needs to be compared to the company’s target
value, which should never be worse than a chance of death of
1x10-4. per annum, the public risk level given in R2P2.
More detailed guidance, particularly of the assessment methods
Office and control room buildings damaged by an explosion and fire at the Bayernoil Refinery on 1 September 2018. Pictures from
Werkfeuerwehr Bayernoil13
Safety practice
Locating a control room near to the plant direct perception. Good system design can assist them to detect
and diagnose problems promptly, giving them the opportunity
The benefits and risks discussed above for remote control rooms to intervene to avoid escalation. These requirements may not be
can be reversed for control rooms close to the plant. However, fully consistent with the ‘normal’ demands for the CRO, which are
there are other issues to consider. more focussed on optimising the process to achieve production
Creating a control room that will protect its occupants from and quality goals, and so need to be carefully considered in the
accidental events can lead to a perception of ‘operating from system design.
a bunker.’ Although nearby, the reinforced structure of the Human Machine Interfaces (HMI) are used by the CRO (and
building may mean the ability to directly perceive the plant is others) to “develop, maintain and use accurate and up-to-date
lost. Providing windows to the outside world becomes more situational awareness of the current, recent past and likely future
complicated (and expensive) meaning they are often not provided state of the system”3. Whilst there can be a lot of discussion about
(or are blocked up as the result of an occupied buildings risk colour schemes, use of symbols and text font, it is the presentation
assessment). This leads to many complaints from CROs who feel of data that makes the greatest contribution. Well-designed
increasingly cut off from the outside world. graphical displays show plant data in ways that is consistent with
Proximity to the plant can encourage CROs to quickly ‘pop human capabilities. They should provide the data the CRO needs
out’ to look at something for themselves, instead of asking a field in a way that they can understand easily without overloading
operator to do it for them. This can lead to the control room being them3. Achieving this requires a thorough understanding of the
unoccupied, with the potential for the CRO to be incapacitated overall system objectives and functions, the tasks performed by
during their brief visit to plant. This is not an issue if an unoccupied the CRO and the information they need to do them. Unfortunately,
control room has been considered in design, but in many cases graphics are often designed simply to show data that is available,
the assumption is for a competent CRO to be present at all times. without a consideration of the CRO’s requirements. Better HMI
Access to fortified control rooms via heavy steel doors and design could have prevented several major accidents including
airlocks can be difficult. Power assistance can reduce the effort Texaco Milford Haven, Esso Longford and BP Texas City1.
required to open and close the doors, ensuring an effective airlock Visibility of data shown on screens and panels depends on
is maintained, but often works slowly. Delays may cause problems viewing distance and angle, size of object (text, symbol etc.), and
if operators need to attend to something urgently on the plant and the person’s eyesight3. Control room designers should have a
the hassle of using the doors may even discourage people with good understanding of how the CROs work when deciding how
legitimate reasons from visiting the control room during normal many screens are required, their size and locations.
situations. Alarms are part of the HMI and are specifically intended to
inform CROs of equipment malfunctions, process deviations
Accident prevention and abnormal conditions. Unfortunately, many systems distract
The CRO has a critical role in ensuring the safety of operations. the CRO with unnecessary and nuisance alarms during normal
They monitor for early signs of problems and intervene to prevent operations, and overload them when things start to go wrong.
escalation. To do this effectively they need to be alert and healthy, Poor alarm management has been identified as a contributory
and supported by well-designed systems. factor in several major accidents including Texaco Milford Haven,
Esso Longford, Cataño oil refinery fire (Puerto Rico) and the toxic
CRO alertness release at the La Porte site in Texas, USA1. Alarm rationalisation
should be an essential activity for any new control room project,
A CRO who is alert and healthy is more likely to detect and
and routinely repeated for operational facilities.
diagnose issues early, reducing the potential for escalation. Given
that many work twelve-hour shifts, including nights, this is not
trivial. The shift pattern and management of hours actually worked Accident response
is critical, but aspects of control room design can also have an It is usually someone in the operating team who will recognise that
effect. a hazardous situation has arisen requiring a prompt and effective
Working conditions in a control room including lighting, response. The situation is unlikely to be obvious at first and the
temperature, air quality and noise will all affect levels of fatigue resources available immediately will be limited5. It is noted that
and stress. Lighting can be very personal, so individuals working in over 50% of the accidents listed in the IChemE summary of major
a control room should have control over their own lighting levels. incidents1 included ‘emergency preparedness’ as a root cause.
Poor air conditioning can contribute to fatigue and other health
issues, and can lead to CROs propping open doors to get fresh air, Identifying a hazard has occurred
which can negate safety and security requirements. Protective system alarms (e.g. fire and gas detection) handled by
Access to welfare facilities including places to prepare and systems independent from the control system may be the first
human factors
consume meals, toilets and rest areas are important because indication that loss of containment has occurred. These do not
“Meal and rest breaks can have a significant effect on CRO tend to suffer with the same problems as control systems but the
performance”3. This assumes that organisational arrangements are way they are displayed to the CRO can have a big impact on how
in place to allow CROs to leave the control room to take breaks, they perceive a developing scenario. During normal operations
which far too often is not the case. single gas detectors may be activated due to faults, routine testing
or small leaks. Identifying a single activation on an alarm list is
Situational awareness quite straightforward. If a large leak ever occurs there will be
CROs achieve situational awareness of plant conditions from multiple detectors being activated and being able to interpret the
control and safety systems, communication with colleagues and pattern can allow the CRO to visualise the flow of the gas cloud
and help them to determine the source and predict the extent of to control room location and architectural design. Subsequent
the hazard. technological developments means that the focus should be on
Being able to see the scene directly can give a better supporting the control room operator (CRO) to perform their
understanding of the issue. CCTV can be very useful for CROs critical role of keeping the plant safe. All design is compromise
in an emergency, and the relatively low cost of systems means and there is no correct solution but there are resources that can
there is little justification for major hazard sites to not have it. help to identify the critical issues and develop optimum solutions.
Whilst a window from the control room is only ever going to This paper has tried to summarise the types of issues that should
provide limited visibility of a plant, the increasing concern of be considered. The following is a (non-exhaustive) list of factors
environmentally caused incidents means that being able to see to consider:
what is happening outside will help the CRO to understand the
impact of heavy rain, strong wind etc. • locate outside the hazardous zone or protect against the
hazard;
Mobilising the appropriate response • make sure the most critical communication is carried out face-
There are some actions that the CRO can take to mitigate an to-face (e.g. shift handover, control of work);
accident. An HMI that allows the CRO to interact with the system • make communication devices readily available and high
quickly and efficiently under all plant conditions can allow them to quality (radios, telephones);
take prompt action. Beyond this the CRO is likely to be directing • support good teamwork within each team with good links
other members of the operating team to take action, activating between teams;
evacuation alarms, mobilising support teams and calling the • allow and encourage legitimate visitors without causing
emergency services. There is a lot to remember and making sure distraction;
emergency response procedures are readily available and fit for
• windows with external views wherever possible;
purpose is critical.
As a scenario develops there may be requirements to formulate • working conditions that enhance alertness;
plans to isolate damaged sections of the plant and vent and/ • lighting that individuals can adjust to suit their personal
or drain process fluids to safe locations. Access to Piping and requirements;
Instrument Diagrams (P&ID), and being able to lay them out • welfare facilities easily accessible;
so that a small group of people can work together is important. • time and cover in each shift for the CRO to take quality breaks
Having a suitable table in the control room, with good lighting away from the control room;
above, should be considered in the control room design.
• HMI graphics designed to show critical information in ways
The control room is sometimes used as the Emergency Control
consistent with human capabilities;
Centre (ECC). It gives the emergency management team visibility
of plant data and allows good communications with the operating • number of ‘normal’ and large screens optimised to show plant
team. However, it is also very distracting for the CRO, who has a overviews and detailed displays;
critical role to play. An adjacent room with visibility into the control • good alarm management so that operators receive early
room may be considered a preferred option. indication that action is required without causing nuisance
and overload;
Allowing CROs to work safely in an emergency • protective systems that provide early warning of hazards;
Although most people on site will evacuate in an emergency, the • CCTV for CROs to visually assess what is going on;
CRO will normally be required to stay in the control room. Power • procedures and supporting information (e.g. P&ID) easily
loss to the site is one common occurrence. Whilst control and accessible with somewhere to lay them out;
safety systems, including associated HMI, are usually supplied
• ECC nearby but separate from the control room;
with Uninterruptible Power Supplies (UPS), control room lights are
not. Designers often consider lighting requirements for evacuating • backup power to all systems including enough lights to
a control room, which ultimately may be required. However, continue working safely in an emergency;
they fail to recognise that the CRO may be required to continue • ventilation systems that prevent ingress of hazardous
working for some time. “Where possible, full lighting should materials.
remain in the control room on power failure. If this is not possible,
the location of lighting units with power backup should take into References
account tasks to be performed during the scenario”3.
Another consideration is the Heating, Ventilation and Air 1. IChemE Safety and Loss Prevention Special Interest Group.
Conditioning (HVAC) system, which “should be capable of Learning lessons from major incidents. (2022)
being operated in recirculation mode if there is the possibility 2. HSE Contract Research Report 432/2002. Human factors
of an abnormal situation resulting in the presence of toxic gas in aspects of remote operation in process plant.
the external environment”3. This feature should be considered 3. EEMUA 201. Control Rooms: A guide to their specification,
as being safety critical and receive appropriate maintenance, design, commissioning, and operation 3rd Edition (2019).
inspection and testing. 4. Brazier, A. Edwards, D. Macleod, F. Skinner, C. Vince, I.
Trevor Kletz Compendium. Elsevier (2021)
Conclusion 5. Brazier, A. Emergency Procedures. Loss Prevention Bulletin
The Flixborough accident highlighted serious concerns related 254 (2017)
Flixborough anniversary
Recollections of Flixborough
A selection from IChemE members
The TV lounge at the University Halls of residence
was full early on Saturday evening, 1st June 1974. As the
closing music of Doctor Who started to play a few people
In 1974 I was a sandwich student in my 3rd year at stood up to leave and prepare for the Saturday night dance
Loughborough University in Chemical Engineering. The course at the University Union. But as the BBC news started to roll
required us to spend a year in industry and I went to Gulf Oil Refining and scenes of a major explosion at Flixborough (a place
Limited in Milford Haven. I remember vividly seeing the accident no one had heard of) were shown, movement towards the
being reported on the news and wondered what caused such a tragic exit stopped and people stared in silence at the screen. (I
disaster. The reporting included visiting the local community – some of also heard about Piper Alpha from the TV in a university
the public had to board up shattered windows and cover their roofs. common room during a “teaching the teachers” course at
The control room staff lost their lives and the plant was virtually Exeter University). Flixborough was keenly discussed by
flattened by the blast. In the ensuing months there was an the chemical engineering students in the following weeks
investigation and I followed this through my colleagues and in the and was in our minds during the following final year of our
press. The findings from the investigation shaped the industry in course. Fifty years ago attitudes were very different, safety
many ways such as better more robust (blast resistant) control room was not a significant issue in the final year design project
designs, better plant layouts, introduction of Management of Change and the requirement for Personnel Protective Equipment
procedures, better control of modifications, and safety in general. (PPE) was limited. Investigations into the devastation at
This was a milestone in history much like Piper Alpha, Pasadena and Flixborough were still continuing when I graduated. After
so on. I have worked in the insurance industry as a risk consultant for starting work, the importance of safety soon became
35 years and we still refer to this as an example of ‘what can happen’ if apparent. Initially in the form of safe methods of work in an
control of work is not managed properly. Industry has benefited from a operational environment and then by the use of rudimentary
learning curve on many occasions – however, we should never rely on risk assessment techniques in a design environment. I
such accidents to make changes to our working methods. have used the basic, and later enhanced tools, of both
Robert Canaway occupational and process safety ever since. For most of that
time, I have conducted insurance risk assessments for a
variety of enthusiastic or disinterested clients. I often have a
sense of déjà vu as issues I identified early in my career are
still common today. Knowledge fades with time and is lost as
one generation gives way to the next. Like rust, it will corrode
if not constantly checked and reinforced.
Doug Scott
I was working in the UK at the time, transitioning to the design office, after ten years in operations at three different sites where I
initially worked as a process operator to becoming a commissioning supervisor on a three-year major oil refinery project.
If I hadn’t seen it all, I thought I had seen most of it – that is the aspect which had the biggest effect on me.
I frankly can’t remember when I heard of the accident, but if I had to guess, my reaction to Flixborough at the time would be more like ‘there
but for the Grace of God’. On reflection, had we, as a society become inured to multiple fatality accidents in heavy industry, mining and air
travel? Remember the Aberfan disaster eight years before and a number of mining multi fatalities and airline crashes. Sure, we moved on
after these by papering over the cracks. But were we doing enough to apply lessons more widely? Almost certainly not.
What did make a massive impact was the enactment of the Health and Safety at Work Act two months after Flixborough. This had been
kicking around in parliament since the 1972 Robens Report, getting nowhere fast during the Conservative Heath government. Did
Flixborough result in HASAWA? Not in itself, but many see it as a catalyst for gaining Royal Assent for the Act two months later when the
Labour/Liberal coalition under Harold Wilson was in power, and I certainly wouldn’t disagree with that. The really important point is that
HASAWA started us on the road to COMAH Safety Cases and that is what really changed things not only in the UK, but within Europe and
other jurisdictions that followed the UK approach to safety. This meant focusing minds by having to demonstrate your safety case to the
regulator and public, and the UK led the way on this. Should we not be celebrating that?
So, did Flixborough change my attitude to safety? Not Flixborough in isolation, but four years later with my operations and design
experience and with some project management under my belt, I was working my first process safety related job, which over the next thirty
years led me all over the world, gaining experience of other cultures, spreading the word, conducting hazard studies,
risk assessments, audits and investigations, and passing on lessons through teaching and coaching. I enjoyed it all, it wasn’t just
a job, for me it was a vocation — the job I was always meant to do.
Undoubtedly one of the major issues coming out of Flixborough was management of change, and this is an area I worked on for almost
all of my subsequent professional career, and thoroughly enjoyed the challenges. Flixborough provides a useful case study, but there are
many others, some more complex and insidious in nature. One of the major areas I was involved in was the development and application
of multi-stage, multi-discipline Project Safety Reviews, later Project HSE Reviews, which are an independent assurance process that all HSE
issues are being properly addressed and closed out by project management. If they didn’t play ball it escalated upstairs, and I’ve got some
interesting stories about that. This programme was eventually linked to the release of capital expenditure at the various stages of a project,
gaining far wider acceptance and adding true value. I led many such reviews and later as a site HSE Manager at two major, and different
sites where major projects were being carried out at those sites and had to respond accordingly.
John Atherton
I was a student in June 1974, just finishing my first year of a PhD at Leeds University. My work included testing concrete samples
for their alumina content, as poorly protected concrete beams made with a high alumina cement had shown a tendency to fail suddenly
and catastrophically, though the failures were fortunately without fatal consequences (in the UK). There is a big read-across to the current
RAAC concrete beam problem! But that meant I already had some awareness of how things could go wrong due
to systemic failures and lack of foresight.
Then on 1st June Flixborough happened and 28 people died in an explosion that destroyed a chemical plant. It was headline news in every
newspaper, on radio and on every TV channel — all three of them! Jackie, my girlfriend (and now Dr Jackie Coates, my wife) was also a
chemist and about to start her PhD at Leeds; her parents lived in North Lincolnshire and she had a motorbike. Her parents’ house was just
4km from the centre of the Flixborough plant and they felt the explosion. Their house was not harmed, though people in the same street
had roofs and conservatories damaged. I visited her at home early that June and she drove us both round the perimeter of the destroyed —
and still gently steaming — Flixborough site. I had visited a chemical plant before, it had seemed vibrant and immensely strong; here was a
site laid waste and its buildings flattened in a moment. For us it was a stark introduction to the potential dangers of
the chemical industry that we would both join in the next 2-3 years.
When I started work in 1976, at a chemical plant in West Yorkshire, the Safety Officer (Gerry Owen) had recently been re-designated Safety
Manager and his department strengthened. I still remember Gerry with great affection and admiration. He was an early advocate of the
then newly founded Loss Prevention Bulletin and I found myself added to the bottom of the circulation list as the magazine
(or possibly illicit photocopies) did the rounds of all the site’s technical and production staff.
Gerry did not have an easy task: from the top of the site distillation columns you could see five working pits and accidents were an accepted
fact of life in the mining industry. That attitude spilled over into our site and many accidents were greeted with a shrug and a claim, via the
Union, for compensation — compensation claimed quite rightly (of course) but it made attitudes harder to shift. Gerry and the production
& engineering management had learned from Flixborough about the dangers of uncontrolled and un-assessed plant modifications and
there was absolute adherence to the newly revised and strengthened “Works Standing Instruction 16b” (known as WSI 16b, or even just “a
16b”), which controlled changes to the physical plant hardware. I was a chemist and I discovered after a while that process
changes were not so well controlled — and neither were changes to the fairly rudimentary control systems on site.
Gerry discovered that I was a physical rather than an organic chemist and that my skill was not in tweaking the last 2% yield out of a reaction
but in understanding the thermochemistry and hazards of our processes. I was given the job of characterising raw materials, products and
processes handled, produced or performed on site, and the company gave me the time and an equipment budget to get on with the work. I
was soon exploding dusts and trying to understand the then fairly new science of plant-scale reaction chemistry – learning from
Gordon Ireland at Ciba-Geigy, Richard Rogers at ICI and Phil Nolan at London South Bank Polytechnic.
I was also given the job of helping to investigate reactions which had not behaved as expected on the plant — accidents and incidents. The
work meant trying to carry out reactions in thermos flasks rather than normal chemist’s glassware and often working outside, sometimes
behind a blast wall, as I allowed (or persuaded) reactions to runaway. Running reactions at a distance, especially if they were intended to
runaway, gave me an interest in both remote sensing of the conditions in the vessel, and in controlling the material flows. I found myself
building systems to remotely control additions, stop and start agitation, and monitor temperature, pressure and other physical parameters
– all from my trusty BBC micro-computer. It was all quite a lot of fun but it also gave me at least some understanding of control systems,
program writing, data acquisition and management, and process characterisation and safety — all of which were of
great importance to me in the rest of my career.
After I’d done ten years in industry, Jackie spotted an advert seeking people join HSE to work on chemical safety and persuaded me
to apply. I must have learned something, as HSE gave me a job as a specialist inspector in process aafety. And then, six years later, the
company I had worked for had a major accident, doing something they had not done before and had not properly assessed. A year after
that I found myself being asked to go back and help the company improve. Despite enjoying HSE enormously the opportunity to work with
my former colleagues was too good to resist and I went back, to spend another 23 years in industry as a safety professional. There were lots
of challenges, not least rapidly adding WSI 16c (for chemical changes) and WSI 16d
(control system changes) to our Management of Change procedures.
It’s probably fair to say I was affected quite a lot by Flixborough and over time it certainly effected big changes in the way industry viewed
process safety. Seeing at first hand what getting it wrong looked like gave me a personal incentive to get it right — or as right as I could
make it. And it affected both of us on that motorbike — Jackie became the Safety Manager for her company and later was the Health,
Safety and Environment Manager for the UK Chemical Industries Association. Together we played a significant role in writing the third
edition of the CIA’s Occupied Buildings Guidance, a publication which was directly driven by the destruction of control rooms,
laboratories and offices at Flixborough. The history of our working lives, in some ways, reflect the changes that Flixborough wrought on the
chemical and process industries — over time changes that have made it considerably better. We both had a working
lifetime of learning from and applying the lessons of the Flixborough tragedy, doing just what Loss Prevention Bulletin
seeks to do in each issue — read on, learn and apply!
Ken Patterson
For me, 1974 came with chartered status four years out from
university. I must try to write this without the benefit of the glorious
hindsight we now have. I had the best start working for ICI Billingham in a great
team studying cyclohexane oxidation (yes, the very same) to cyclohexanone
and cyclohexanol (KA), intermediates in the production of nylon. ICI had two
KA Plants, a smaller one at Billingham which used an ICI process and a larger
one at Wilton which was an SDC (Scientific Design Company) plant which I
understand to be like the one at Flixborough. At the time of Flixborough I had
been transferred to be the sole Chemical Engineer on a similar team studying
the oxidation of paraxylene (also capable of a big bang) to terephthalic acid – a
polyester intermediate. During this period, I was seconded to Davies works
Wilton to help with a plant start-up and while there the news of Flixborough
came through. As was our custom in those days our lunch breaks were spent
with a run followed by a nutritious pie and a pint. This group of smart, ambitious
young process engineers pondered as information emerged and we wondered
why the bang was so big and what could be done to minimise the chance of
what appeared to us at the time to be a one in a million occurrence. Trevor Kletz
was Petrochemicals Safety Advisor at the time and was making noises. So, we
all benefited from his HAZOP training and attendance at many workshops.
I later hosted Trevor for a short time during a trip to New Zealand. And thus
began a lifelong relationship with process safety, a word that did not exist at the
time, and carried on and developed through my years in plant management,
engineering management and eventually process safety consultancy. Over time
I have clarified my views on the role of HAZOP, coming to the realisation that
when designing, engineers are focused on making a process work, not fail. It
is the latter mindset which is required by the team when conducting a HAZOP
study and the leader must ensure it is maintained. The company I started, Safety
Solutions Ltd now run by my son, also a Chemical Engineer and Fellow of the
Institution is now expanding beyond New Zealand and promoting all aspects
of process safety. If you will excuse an awful pun. Flixborough was the spark
that started it all and I will never forget the closing comment from the BBC
documentary on Flixborough. “Just one slip and the abyss”.
Colin Feltoe
I was a second year engineering student at the time of the incident and do not recall hearing about it then.
However, when I started work in Santos Australia, references to the Flixborough incident and learnings from the incident
had made it to many safety related and other courses. Subsequently as I progressed in Santos and took charge of process
safety, Flixborough always got a mention whenever and wherever Management of Change (MOC) was discussed. A
second major learning from Flixborough, as I recall, related to impact of unconfined vapour cloud explosions (VCE) on
occupied buildings. This second aspect in my view did not get the attention it deserved. When I say attention I mean
developing management systems as well as technical standards on the topic. In my work and also in my interactions with
ISC and CCPS, we tried to revisit learnings from major incidents including Flixborough and fill gaps if or when found. My
take on where we stood, and may be still stand, regarding the two major learnings from Flixborough are as follows:
• MOC: While MOC processes are quite well established in most major organisations, what is not often clear
is what “what qualifies as change” and “who is responsible for recognising change” and whether or not well
equipped to call a “change” when it happens. The point I am trying to make is that once a change is recognised
, due processes get initiated to respond to it, but it is the very recognition of “change” that is challenging and
industries must continue to develop this recognition aspect on a continuing basis.
• VCE: Unconfined VCE modelling is quite advanced and does get called upon more often than not for new
projects and developments. Existing facilities however may not draw the same attention. Secondly technical
standards for temporary buildings were not developed enough to address capability of withstanding a VCE
overpressure, at the time of my leaving the industry six years ago. I am not sure if much has changed. And
although a priority of temporary buildings, the consequence analysis is equally important for permanent
buildings that may not have undergone a review for overpressures resulting from VCE.
I have focussed on what I took away from the incident and how it shaped me as well
as what I think might still be out there requiring more attention.
Shekharipuram Sreedhar
I was a 13-year-old teenager with no idea that my working life would be spent
as a chemical engineer in the process industries, but I still have vivid memories
of the Flixborough disaster and the lessons from it have been with me ever
since. I remember hearing about the accident very soon after it happened; the
football results on the TV were interrupted to announce the tragedy and I recall
seeing the first aerial footage of the scale of the explosion, however, growing
up on Tyneside, I had no idea where Flixborough was.
I recollect the news of the incident as it happened on my birthday, 1st June 1974. I was sitting for my A Level
exams after completing a one year crushed course for A Levels. This is because the pre-university education that I had
in Bangladesh was not up to A Level standard. Although it was horrific, the news of the accident did not make as big an
impact as I think it should have done. This is probably because the chemical industry (and industry in general) was used
to having frequent accidents at that time. Later that year, in the autumn, I started a chemical engineering degree course
at Imperial College. During my study at Imperial, the significance of Flixborough was brought to me very extensively in
safety lectures by Dr. Napier. If I remember correctly, Dr. Napier was in the Flixborough investigation team. I graduated in
1977 and joined the Humber Refinery of Conoco Ltd. (now Phillips 66). The refinery is in North Lincolnshire (then South
Humberside) and is only about 30 miles from Flixborough. It would be appropriate to say that process safety, as we know
it now, was not there. The refinery was proud of its safety record but safety was measured in terms of personal injuries as
it was throughout the industry. Process safety was part of engineering judgement as evaluated by experienced engineers
and operators. Elements of process safety e.g. HAZOP was introduced in the 1980s. CIMAH regulations required the
refinery to produce safety reports justifying our safe operations. This was, of course, intensified by COMAH regulations
(as an aftermath of Seveso and Bhopal incidents).
After doing various assignments in technical services, operations and projects, I was definitely attracted to process safety.
It was almost a natural progression and it became clear to me that experienced chemical engineers like myself must
become guardians of process safety and question everything to maintain plant integrity. This personal mission became
more enhanced after our gas plant explosion in 2001. The Nypro Flixborough plant was rebuilt and I had an opportunity
to visit the plant on an IChemE organised plant tour. It was a model plant with every structure fireproofed and automatic
water sprinklers everywhere. Unfortunately, the plant was not profitable and was closed down.
S. Mohammad Ali
Safety leader
knowledge and
through open, frank discussion and pooling our different
competence
Ken is a member of the Industry Safety Steering Group
monitoring industry’s progress in implementing the perspectives that we could best achieve that goal. It led
Building Regulation and Fire Safety review post-Grenfell. to industry becoming more self-disciplined and holding
He also chaired the industry/regulator task force in the ourselves more to account and it led to a more mature and
wake of the Buncefield Terminal explosion. Ken has a track collaborative relationship with the regulator. It led to leaders
record of managing change and business turnarounds across organisations stepping up and holding themselves to
engineering
and design
based on developing a clear strategic direction and account. When industry works together, when regulators
building organisational capability. work together and then when industry and regulators work
together then transformational change can happen. And that
is magic!
At what stage in your career journey did the The success of the Buncefield Standard Task Group in
systems and
procedures
importance of leadership in process safety working together to identify, develop and deliver real change
was subsequently continued and built on by the Process
become clear to you?
Safety Leadership Group. The incidents at Buncefield —
The moment I think the absolute and truly critical importance and also at Texas City — highlighted and re-emphasised
of leadership for me was triggered by Buncefield. I saw the the critical importance of leadership in preventing major
profound impact that leadership can have not only in how we incidents. The PSSG went on to define what good leadership
manage major hazards within an organisation but also across in managing major hazards looks like. That work on
a nation. leadership had a resounding impact on the UK process
Buncefield was a shock to the industry, it was a shock to the industries and is now embedded in the regulatory framework.
regulator and most importantly a shock to the public. See Principles of Process Safety Leadership: http://www.p-
A Major Incident Investigation Board was set up to s-f2.org.uk/wp-content/uploads/PSLG-Principles-1.pdf
identify what went wrong, but it took time as much of the I think the leadership which drove the Buncefield Task
evidence had been destroyed. In the meantime, pressure was Force went beyond the technical and operational and facility
mounting on all parties to do something. changes that were implemented (real and important as they
The crisis around Buncefield changed the nature of the were). It demonstrated a new level of maturity in our ways
interaction between regulators and the regulated in the UK. of working and it paved the way for the working together on
Up until then, developments progressed like a game of tennis leadership. Better outcomes in turn build trust and credibility
with one party proposing change which would be rebuffed between regulators and regulated, and created a virtuous
with counter proposals from the other side. Solutions and spiral, which manifests itself today in the COMAH Strategic
Forum, which was set up to bring the Competent Authorities on a journey that has helped me grow those leadership skills
and the “chemical sector” together to discuss and resolve and I am still learning today.
matters of strategic importance in the management of major My leadership journey started with trying to understand
hazards. what a leader really is. There are thousands of books written
As I said back in 2007 ….. about thermodynamics and they all say the same/similar
things. There are probably even more books written about
“How industry responds to incidents such as Buncefield,
leadership and they all seem to say something different.
and how the regulators respond on behalf of the public is a
measure of our society. A decisive and dynamic response So what do I mean by being a leader and how is that
with all parties co-operating is the product of a democratic different from being a manager? Managers are great at moving
and advanced society...” from A to B, but it is leaders that define what B is and it is that
vision in how it is created, defined, shared and owned that can
And the work of leaders in making that happen is pivotal. transform organisations.
For me it was always about what I called “visible leadership”,
Was there a key event or person that had a about creating a compelling vision, then going on to build
big impact on shaping your interest in process a guiding coalition to drive that change. To excite, energise
safety. What happened and what did you and empower people and provide the support and resources
they need……..and then to step back and let them deliver, but
learn? always being available to help, support and encourage.
The key event was Flixborough, and the key person was Trevor It was always coupled with high and well understood
Kletz. standards in which to instil the belief and confidence that we
Flixborough occurred in 1974 while I was at university can do better and be different, to recognise that you get the
and the following year I started my career with Shell at performance you demand and to hold especially yourself and
their refinery in Essex. After a year working in crude oil then others accountable.
distillation, I joined the Major Projects group which had I always remember a story told by a colleague of mine. He
a big focus on managing major hazards as a result of had just become site manager of a major chemical complex
Flixborough. I remember being busy with reviewing the for the first time. There was an incident which released a large
management of change, but also becoming one of the first cloud of flammable gas that flowed past their administration
technologists to be sent off to learn from ICI about HAZOPs centre full of people. Fortunately, the gas cloud dissipated,
etc. and put that into practice. and no one was injured. His first reaction was “how could they
What a life changing experience that was! To meet Trevor have done this? how could they have let this happen?” He was
Kletz and his team and be exposed to the eye-watering cross and angry. The following morning when he woke up,
array of potentialities for disaster and to then grasp how he realised how wrong he was. The question he should have
that potential chaos could be converted into an orderly and asked himself first was “how could I have done this? how could
structured management of the risks. To understand how I have let this happen?”
chronic unease did not have to be disabling but motivating. As a leader, the organisation looks to you. Your responses
And of course, to grasp the idea of inherent safety — “What or lack of response send powerful signals which shape their
you don’t have, can’t leak”. behaviour and attitudes, influence decisions and priorities. I
I guess the biggest lesson though was about personal always look to myself first and what it is that I did (or didn’t do)
conduct. About not asking for permission, but just doing that contributed to any incident or shortcoming. That has led
what was right. As a result, during my career, I have had to to a fundamentally different approach and outlook in managing
ask for lots of forgiveness after the event, but I have never major hazards and which has helped me on my leadership
regretted choosing that path. journey.
Trevor was a true pathfinder and has inspired me and I guess the final comment though is about changing the
generations of other chemical engineers to grasp the nature of conversations.
difficult challenges of addressing those major accident I believe that more diverse inputs leads to better
hazards that can kill, maim, or wound people, do untold conversations which lead to better analysis and better
damage to the environment, generate huge economic cost conclusions, better decisions and ultimately better outcomes.
and disruption, and destroy communities and businesses. And that requires you to be open to listening, hearing and
He shone a light on how those events can be mapped, understanding others’ views especially if they are different
mitigated, and managed, and those lessons and insight to yours. And if you show that you are interested in others’
are now being grasped beyond the process industries, for views and you respect them then a rich cascade of new and
example in the building sector post-Grenfell. fresh inputs can emerge that helps address those difficult and
sometimes intractable process safety questions.
How do you, as a leader, ensure that People make the difference and creating an environment in
your commitment to process safety is which they can deliver their best is essential.
communicated, embraced and correctly
applied by your teams? As a leader, how do you evaluate and
Firstly I think you have to be yourself and demonstrate your
balance risk?
personal commitment to managing major hazards. I have been There are three questions that are always top of mind. Do we
Visible Leadership
• create the compelling vision
• build the guiding colation
• excite, energise and empower
• support, resource, stand back and encourage
understand the risks we are managing? Do we know how to also to look at the learnings from other incidents that might be
manage and mitigate those risks? And have we reduced them occurring in your sector and asking whether the lessons apply
to as low as reasonably practicable (and can we demonstrate to you. And ultimately, what did the Grenfell fire or the Boeing
that)? 737 Max crashes mean for your business.
There are processes, tools and techniques that help us, but Underpinning all this is a culture of “chronic unease’ that
central is an open and transparent discussion in which issues goes beyond compliance and that questions what more can
are shared and in which we respond strongly to weak signals. and should be done.
It requires a culture in which the raising of problems, concerns And those questions apply to both the management
and uncertainties are welcomed. of specific risks but also the steps in the improvement
And rather than looking at specific risk, I would look at how engine which underpin the culture in the organisation. The
we establish within an organisation that right culture in which journey from “unconscious incompetence” to “conscious
this occurs. This starts with the journey that an organisation incompetence” relates to a dependent culture where external
goes on starting with “unconsciously incompetence” in regulatory standards often dominate and where you feel you
which the sky is blue, grass is green, and everything appears are doing it because you have to. The next step in becoming
OK — and then we have an incident and we realise that we “consciously competent” is driven by an understanding
are “unconsciously incompetent”. Hopefully, we realise this
before an incident occurs by taking measurements, trending
those measurements and benchmarking the results to see Improvement Engine
whether we have a performance problem.
Having realised we are “consciously incompetent’, we then
UNCONSCIOUS
go on to implement the systems, processes and procedures COMPETENCE
that if followed will lead us to become demonstrably CONSCIOUS CULTURE
COMPETENCE
Audit, Learning from others
“consciously competent” and have passed our “driving test”! Values, Behaviours
Understanding hazards
■ Advanced Process Safety Considerations for
Hydrogen Projects
■ Pressure Systems
■ What Engineers Need to Know About
Hydrogen Safety
Human factors
■ Human Factors in the Chemical and Process Industries
IN-COMPANY
on-site or online.
www.icheme.org/process-safety-training