Process Safety and Environmental Protection 9 7 ( 2 0 1 5 ) 109–115

Contents lists available at ScienceDirect

Process Safety and Environmental Protection

journal homepage: www.elsevier.com/locate/psep

Causes and consequences of thermal runaway

incidents—Will they ever be avoided?

Rim Saada, Dipesh Patel 1 , Basudeb Saha ∗

Centre for Green Process Engineering, School of Engineering, London South Bank University, 103 Borough Road,
London SE1 0AA, UK

a r t i c l e i n f o a b s t r a c t

Article history: A study of runaway incidents involving thermal chemical reactions in the UK over the past
Received 19 September 2014 25 years (1988–2013) has been carried out. The objective of this study is to determine possi-
Received in revised form 9 January ble causes of thermal runaway incidents. A statistical analysis of the underlying problems
2015 that led to thermal runaway incidents has been provided. A comparison of the current study
Accepted 4 February 2015 on thermal runaway incidents with those identified prior to 1988 has been carried out. This
Available online 12 February 2015 study clearly shows that lessons have not been learnt from thermal runaway incidents
caused by operator errors, management failures and lack of organised operating proce-
Keywords: dures. These factors have been the possible causes of about 77% of all the thermal runaway
Inherent safety incidents analysed in this study. The number of fatalities and injuries as a result of thermal
Hazard identification runaway incidents has increased by ∼325% and ∼279%, respectively, in the last 25 years even
Health and environment though the number of incidents was significantly less. On the basis of this analysis, several
Process design recommendations have been proposed that could help to minimise the risks associated with
Risk assessment any thermal runaway incidents in the future.
Thermal runaway reactions © 2015 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

1. Introduction factors. This increases the possibility of thermal runaway as

Thermal runaway reactions are characterised by progressive
increase in the rate of heat generation, temperature and
pressure (Barton and Nolan, 1991). Heat generation increases
exponentially with an increase in the system temperature. It
also increases due to other factors including the lack of process
control and failure to cool the reaction system. An increase in
the system pressure could occur due to vapourisation of some
of the components in the reaction mixture and decomposition
of some of the gaseous products at the elevated temperatures.
In batch operations, the rate of reaction and production
are controlled by maintaining the amounts of the reactants,
solvents, catalyst and non-reacting chemicals charged to a
reactor. Generic batch reactors are usually used for different
types of chemical reactions in an industry rather than being
specifically designed for a particular reaction due to economic
it could be ignored that the heat of some of the reactions
may exceed the existing cooling capacity of the reactor. Semi-
batch operations tend to be used for specific unit processes
that are inherently hazardous such as halogenation, nitration,
polymerisation and sulphonation reactions. Similarly, ther-
mal runaway could be a particular problem in unsteady-state
batch reactors as reported by Barton and Rogers (1997). This is
mainly due to a difficulty in specifying the design, operation
and control of a stirred reactor, heating/cooling coils, reflux
facilities and emergency relief venting, which would require a
systematic approach.
Chemical reactions conducted in batch reactors may get
out of control due to other reasons including change in an
operating conditions and usage of inappropriate materials
(Gillard, 1998). Since batch operations involving exothermic
reactions are common in industries, precautionary measures

∗
Corresponding author. Tel.: +44 0 20 7815 7190; fax: +44 0 20 7815 7699.
E-mail address: b.saha@lsbu.ac.uk (B. Saha).
1
Present address: School of Engineering, University of Hull, Cottingham Road, Hull HU6 7RX, UK.
http://dx.doi.org/10.1016/j.psep.2015.02.005
0957-5820/© 2015 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
110 Process Safety and Environmental Protection 9 7 ( 2 0 1 5 ) 109–115
Nomenclature
HSE Health and Safety Executive
H enthalpy change (J)
LPB Loss Prevention Bulletin
SDS safety data sheet
SOPs standard operating procedures
are necessary to minimise the risks associated with ther-
mal runaway. Some reactions may constitute reaction hazards
due to the complexity of the reaction, e.g., nitration reac-
tions are considered to be the most destructive reactions in
the present chemical industries as it involves exotherms and
heat-sensitive products (Mannan, 2012).
Good control and hazards associated with the chemical
reactions are essential aspects of chemical manufacturing.
Therefore, the aim of this work was to study runaway inci-
dents for the last 25 years, i.e. 1988–2013 and to compare it
with the findings reported by Barton and Nolan (1991) for the
thermal runaway reaction incidents occurred between 1962
and 1987. It is envisaged that lessons could be learned from
the recently reported incidents and those identified before
1988. The findings in this article include a detailed analysis
of the type of industries, unit processes, causes and conse-
quences. It is expected that the information provided in this
paper would incorporate the changes within numerous indus-
tries to minimise the risks associated with thermal runaway
reactions.
2. Incident statistics
2.1. Case histories pre-1988 (1962–1987)
Barton and Nolan (1991) studied case histories for industrial
incidents in batch reactors involving thermal runaway reac-
tions from 1962 to 1987. They reported that the main causes of
thermal runaway incidents were process chemistry and plant
design and operation. The lack of understanding of reaction
chemistry, temperature control and mischarging of reactants
were the main causes for 20%, 19% and 21% of the incidents,
respectively.
2.2. Case histories post-1987 (1988–2013)
Possible unit processes have been identified for 30 incidents
involved in thermal runaway reactions. However, it was not
possible to obtain detailed analysis of every incident due to
lack of information in the public domain. Nine unit processes
that were involved in the incidents are shown in Table 1. It is
evident from Table 1 that one third of the incidents have been
contributed by polymerisation process followed by decompo-
sition process (13.3%).
2.3. Causes of the incidents studied
The possible causes that lead to an overheating and eventual
thermal runaway for 30 incidents are classified under “Tech-
nical and Physical Causes” and “Human and Organisational
Causes”. These causes are explained in detail in Sections 2.3.1
and 2.3.2, respectively.
2.3.1. Technical and physical causes
Most of the incidents reported in this study have multiple
causes, i.e. some of the incidents initiated due to a technical
or physical cause were extended to a thermal runaway due
to a human cause (LPB, 1995). Fig. 1 represents the number
of incidents occurred in the last 25 years with technical and
physical causes. It is evident from Fig. 1 that mischarging the
reactor has contributed to five incidents. This includes addi-
tion of incorrect amount of chemicals and charging chemicals
or catalysts in an inappropriate order. In one incident (LPB,
1994b) the root cause of the incident was an overcharge of
vinyl acetate during the monomer emulsion make-up for the
batch. This was accompanied by inadequate management of
a change in the operating procedures.
In this study, four incidents have been caused due to the
failure of agitator device (Fishwick, 2004, 2008; LPB, 1993b,
1994a). In one incident the agitator seal ring blew due to an
increase in pressure in the reactor (LPB, 1993b). In another inci-
dent, the agitator failed causing a cooling at an early stage in
the nitration reaction (LPB, 1994a), whereas, in one incident
the propeller failed to start and therefore caused a build-up of
the temperature in the upper part of the reactor content lead-
ing to a thermal runaway incident (Fishwick, 2004). During an
incident, the agitator started to slip when a fault occurred in
the connection or linkage between the agitator and its drive
mechanism and hence the agitator lagged behind and failed
to stir properly (Fishwick, 2008).
Trace quantities of impurities could greatly affect a chem-
ical reaction either by reducing the rate of reaction (Ahmad
and Baloch, 2007) or increasing the rate of the reaction (Gustin,
2002). During an incident traces of water were retained after
washing the reactor that caused an unexpected thermal
runaway reaction (Kletz, 1995). The contamination of pure
chemicals by trace impurities could lower their thermal stabil-
ity to a larger extent and cause an unexpected decomposition
under normal process conditions (Gustin, 2002). In this study,
three incidents were caused as a result of contamination
that contributed to thermal runaway and has been reported
under ‘Quality Control’ sub-heading of ‘Technical and Physi-
cal Causes’ (Fig. 1). One incident could have been prevented
if the solvent was not contaminated with acetic acid (Kletz,
1991; Van Reijendam et al., 1992).
Reactor design and temperature control are important in
a chemical plant to ensure safety of both the operators and
the plant. Temperature inside the reactor may deviate due
to changes in the heat input/output, heat transfer, pressure
and due to thermal lags and hot spots. In this study, four inci-
dents occurred due to a poor plant design (Lindley, 2001; LPB,
1993b, 1994a, 1996) and five other incidents were caused as a
result of failure in the process control (FACTS, 1988, 1990, 1991,
1994; LPB, 1993b). Hence, it is vital that mechanical design of
all reactors is completed to highest standard. In one incident
(LPB, 1996), the reactor failed due to a stress rupture because
the reactor was allowed to operate at excessive temperatures
beyond its safe operating envelope. This could have been pre-
vented if operating limits were checked beforehand by the
operator or the reactor could have been designed to withstand
such excessive conditions.
Malfunction, technical failure, venting, leakage, loss of pro-
cess control and power cut together have possibly contributed
to sixteen incidents. All these causes could have been avoided
in the process industry if the operator knows beforehand how
to deal with a particular situation.
 Process Safety and Environmental Protection 9 7 ( 2 0 1 5 ) 109–115 111

Table 1 – Number of incidents in a specified unit process during 1988–2013.

Unit Process Total incidents Incidents (%) References

Cracking 1 3.3 LPB (1992)

Decomposition 4 13.3 Fishwick (2004), LPB (1993a), Pinsky et al. (1990), Sachdev and
Todd (2005)
Ethoxylation 2 6.7 Gakhar and Carr (2003), Lindley (2001)
Hydrogenation 1 3.3 Van Reijendam et al. (1992)
Nitration 1 3.3 LPB (1994a)
Polymerisation 10 33.3 Fishwick (2008), Gillard (1998), Kletz (1995), LPB (1990, 1993b,c,
1994b, 1995), Shu and Wang (2003), Whitmore et al. (1993)
Refinery 2 6.7 FACTS (1994, 1990)
Storage 1 3.3 FACTS (1992a)
Transportation 2 6.7 FACTS (1999, 2006)
Others 6 20 FACTS (1988, 1989, 1991, 1992b,c), LPB (1996)

2.3.2. Human and organisational causes
A number of human and organisational causes have been con-
tributed to thermal runaway incidents in the last 25 years as
illustrated in Fig. 1. It is apparent from Fig. 1 that operator
errors have contributed to the majority of the incidents (13
incidents). It was not possible to identify the exact error due
to the lack of information in four incidents (FACTS, 1988, 1989,
1990, 1992b). The lack of understanding of the chemistry by
the operator was found to be responsible for thermal run-
away incident in one instance (LPB, 1992). In another incident,
the operator failed to realise that using butene instead of a
mixture of butene with other compounds was a modification
to the existing process that should have been systematically
approached (Kletz, 1995). Similarly, in another incident the
operator thought that water was completely removed from
the process unit and did not double check whether the water
was removed or not (Kletz, 1991; Van Reijendam et al., 1992).
Operating outside the safety limits of the reactor vessel has
contributed to one incident where the responsibility rested
with the operator and not with the technical advisor (LPB,
1996). Inadequate training (Sachdev and Todd, 2005), absence
of supervision (Pinsky et al., 1990), an increase in work load
(LPB, 1994a), failure to follow standard operating procedures
(SOPs) (Lindley, 2001) and incorrect opening/closing of the
valves (LPB, 1993b) have each contributed to one incident in
the current study. The findings in this study are in agreement
with a study carried out by Mannan (2012). It was reported that
large losses in the process industries were caused due to oper-
ator errors that contributed about 17% of the causes (Mannan,
2012).
Appropriate management and supervision are essential for
smooth running of a chemical plant. Poor management was
blamed for an incident in 2008 where monomer mixtures were
left static for prolonged period and the electrical installation
was not fully repaired from the previous day thus causing a
thermal runaway (Fishwick, 2008).
Organised procedures must be followed with safe practice
at all times when dealing with chemical reactions. SOPs
should be checked and amended on regular basis in order to
minimise the risks and hazards associated with a plant opera-
tion. Failure to follow these procedures may result in failure to
control the reaction. Managers should take into account that
adequate staffing could help in identifying early signs of ther-
mal runaway and could possibly act to prevent it. Enough staff

Fig. 1 – Possible causes of 30 incidents reported during 1988–2013.

112
Paint and varnish
Food and flavourings
Agricultural chemicals
Dyestuffs
Manufacturing Industries
Metallurgy and metal processing
Refinery
Inorganics
Laboratory
Pharmaceuticals
Miscellaneous
Plastics, rubbers and resins
Fine, intermediate and heavy
organics
0 20
Fig. 2 – Comparison of the incidents in different
manufacturing industries as reported by Barton and Nolan
(1991) with the current research.
should be available at all times to operate and monitor a given
chemical plant, however, many operators are made redun-
dant in order to cut cost and generate more profit. Educating
managers, supervisors and operators about the safe operation
of the chemical plant is essential in order to avoid incidents
and minimise losses (Fishwick, 2012). Cleaning of chemical
reactors and equipment should be conducted appropriately
to ensure that it is suitable for the process materials.
2.4. Industries involved
Batch reactors are ubiquitous in the chemical industry due to
their convenience and flexibility (Vasanthi et al., 2012). How-
ever, the pattern of incidents shows preponderance to certain
specific industries as reflected in Fig. 2. Fig. 2 shows that ∼43%
of the incidents occurred in fine, intermediates and heavy
organics industries, while ∼30% of the incidents took place
in plastics, rubbers and resins industries.
2.5. Consequences
The results from runaway incidents ranged from a simple
foam-over of the reaction mass to a substantial increase in
temperature and pressure leading to a violent loss of contain-
ment and a release of large quantities of flammable, hazardous
and toxic materials. In one incident 45 kg of vinyl acetate was
released to the atmosphere (LPB, 1994b). In a couple of cases,
flammable materials were released that caused fire and a sec-
ondary explosion followed (Kletz, 1991; LPB, 1994b). This study
realised that 17 fatalities and 311 injuries occurred during the
period 1988–2013 (see Table 2). A large number of the injuries
occurred due to release of toxic fumes and their subsequent
inhalation (Gakhar and Carr, 2003). These incidents generally
occurred when the operators were attempting to regain con-
trol of the reaction.
Process Safety and Environmental Protection 9 7 ( 2 0 1 5 ) 109–115
0.00 Current Research Table 2 – Comparison of deaths and injuries reported by
2.65
Barton and Nolan (1991) Barton and Nolan (1991) with the current study.
0.00
2.65 Field Barton and Current %Increase
0.00 Nolan (1991) research
2.65
0.00 Deaths reported 4 17 325
6.88 Injuries reported 82 311 279
0.00
6.88
3.33 3. Discussions
0.00
6.67 3.1. Statistical analysis of the current work
0.00
10.00
0.00 In this study, 30 incidents have been examined. The possible
6.67 causes of the runaway incidents from this study have been
6.88
compared with those studied by Barton and Nolan (1991) and
0.00
12.17
they are as follows:
30.00
It is apparent that a similar trend occurred in the technical
21.69 and physical causes of runaway incidents for both pre-1988
43.33 and post-1987 studies. Barton and Nolan (1991) reported that
37.57
the main physical and technical causes of the incidents were
40 60 80 100 mischarging (35 incidents), reaction chemistry/development
Incidents (%)
not conducted beforehand (34 incidents) and agitation (17 inci-
dents). Similar trends could be seen for the current study
where the possible physical and technical causes are mis-
charging (5 incidents), plant design (4 incidents) and agitation
(4 incidents). One of the similarities in the trend was the
mechanical failure (such as stirrer blades shearing off), which
contributed to three incidents for pre-1988 and post-1987 stud-
ies. On the other hand, failure due to temperature (i.e. loss of
process control) decreased from thirty two incidents occurred
in the pre-1988 study to five incidents for the current research.
Leakage, venting and power loss have contributed to six inci-
dents. This study has reported six incidents due to technical
failures, but not enough information is available to identify
the actual reasons for these failures.
Human and organisational errors are one of the possible
causes of the incidents in the chemical industries as shown
in Fig. 1. The operators have contributed to thirteen incidents
studied in this work as compared to eleven incidents reported
for pre-1988 study. The major reduction was found to be in
the maintenance department, where twenty five incidents
were reported between 1962 and 1987 as compared to only
one incident in the last 25 years. Incorrect labelling, clean-
ing, inappropriate staffing are all new causes which have not
been reported in pre-1988 study. Hence, it could be concluded
that apart from runaway caused by maintenance and process
control, lessons have not been learnt from previous incidents.
Fig. 2 shows thermal runaway incidents in various man-
ufacturing industries for pre-1988 study and the current
research. It can be seen from Fig. 2 that fine, intermediate and
heavy organics followed by plastics, rubbers and resins are the
main industries involved in the thermal runaway incidents
as reported by both pre-1988 study and the current research
(1988–2013). Fig. 2 highlights that there has been no relief for
industries associated with plastics, rubbers and resins, where
the thermal runaway incidents have increased by ∼8%. Sim-
ilarly, the thermal runaway incidents across the industries
associated with fine, intermediate and heavy organics have
increased to ∼6% in the last 25 years. An increase in incidents
in these industries could be due to more sophisticated nature
of the chemical reactions.
It can be seen from Table 2 that the number of deaths
and injuries are a lot higher in this study as compared to
pre-1988 study even though the previous study had covered
 Process Safety and Environmental Protection 9 7 ( 2 0 1 5 ) 109–115
189 incidents as compared to only 30 incidents that are cov-
ered in this work. The incidents occurred in the last 25 years
must have been more serious in order for the number of fatal-
ities to be this high. In the past, there were many operators
working on a site, whose main job was to look and identify
for any sign of incidents. Currently, the situation in the indus-
try has been quite different. Many operators have been made
redundant in order to minimise cost and generate more profit.
Managers have failed to take into account the vital role the
operators used to play in identifying early signs of runaway
incidents. This could be one of the reasons why explosions
after 1987 were more serious and caused greater fatalities.
Managers should be aware that safe operation and sustainable
business success could not be separated (Travers and Smith,
2012). On the other hand, some types of the earlier runaway
incidents reported by Barton and Nolan (1991) do not occur
now because engineers, operators and managers could have
taken necessary preventive measures to avoid such causes of
thermal runaway incidents as reported by Barton and Nolan
(1991).
3.2. Lessons learnt
The analysis indicates that incidents occurred due to lack
of basic understanding of the process chemistry, inadequate
engineering design for heat transfer, inadequate control sys-
tem, no safety back-up systems (including venting) and
inadequate operational procedures, including training.
In order to deal with a hazard, it is necessary to first iden-
tify it using various hazard identification techniques as shown
in Fig. 3. Hazards should be checked on how likely they could
occur and how serious the consequences could be. It is evident
that in order to avoid conditions for runaway arising in the
first place, it is necessary to have the knowledge of both the
chemistry, associated thermo chemistry of the desired reac-
tions, thermal stability and physical properties of reactants,
intermediates and products. Some of this important informa-
tion could be obtained from the literature, safety data sheet
Project Stages
Chemicals testing
Impurities testing
Research and Development
Reactions testing
Reactivity and explosibility
Hazard studies
Pre-Design Insurance assessments
Pilot plant
Failure modes and effects analysis
HAZOP studies
Design Process design checks
Risk and reliability assessments
Task analysis and operating instructions
Checks against design
Emergency planning
Commissioning Inspection examination monitoring
Plant safety audits
Pre-startup and safety review
Condition monitoring
Corrosion
Operation
Plant safety audits
Regular Inspection
Fig. 3 – Hazard identification techniques at different
process stages.
113
(SDS) or from computer-based modelling of reactions. A vari-
ety of laboratory techniques such as adiabatic calorimetry and
accelerating rate calorimetry are available for use to acquire
knowledge about the thermal behaviour of different reaction
components.
3.3. Protective measures
The Health and Safety Executive (HSE) reported that the basis
of safety of a chemical reaction is the combination of measures
that are relied upon to ensure safe operation (HSE, 1998). The
measures that are chosen for a particular case will depend
on a number of factors including the probability of runaway
reactions occurring, the applicability of various methods and
the compatibility of the measures with the plant operation.
In practice, it is not possible to eliminate all hazards by
inherently safer methods, but additional control measures
could be added to minimise the risk and back up the con-
trol measures with protection, such as process control along
with the use of inhibitors, emergency relief systems, quench-
ing or containment. Such combination of methods is common.
Runaway incidents may also affect the environment. Hence,
it is also important to consider whether measures taken are
adequate to comply with the environmental regulations.
3.4. Process control
According to the HSE, process control includes the use of sen-
sors, alarms, trips and other control systems that either take
automatic action or allow for manual intervention to prevent
the conditions for uncontrolled reactions. The specification
of such measures requires a thorough understanding of the
chemical process involved, especially the limits of safe opera-
tion (HSE, 1998).
3.5. Process control and inhibition
In batch operations, a failure in the cooling and stirrer drives
are problematic since the heat produced by the reaction can-
not always be removed sufficiently. In order to avoid the
release of the reaction mixture into the atmosphere, an
inhibiting agent could be injected as a first line of defence. A
chemical inhibitor is usually applicable only to intermolecu-
lar reactions rather than decomposition reactions (Barton and
Rogers, 1997). In the polymerisation of vinyl compound when
a runaway is detected, reaction inhibitors, e.g., tert-butyl cat-
echol could be fed to stop the reaction by removing the free
radicals that propagate it. The right choice of the inhibiting
agent for each specific reaction system and the rapid homo-
geneous mixing of the inhibiting material with the reaction
mixture is a basic requirement for a successful termination of
the reaction. Laboratory studies should be carried out in order
to determine the inhibitor concentration, rate of addition and
the temperature of the chemical reaction. The reliability of the
inhibition system must be quantified and demonstrated to be
acceptable if inhibition is used as the ultimate basis of safety.
3.6. Process control and emergency relief system
A key decision in selecting safety measures is whether or not
to fit the reactor with a pressure relief system designed specif-
ically to protect against the hazard from a runaway chemical
reaction. Pressure relief system is usually fitted to the
114 Process Safety and Environmental Protection 9 7 ( 2 0 1 5 ) 109–115
chemical reactor in order to cope with an exothermic runaway
reaction or if the reactor is engulfed in fire (Hellemans, 2009).
One of the biggest problems in sizing vents is the avail-
ability and accuracy of physical property data for the reaction
components. It is a good practice when sizing a relief system
to utilise several design methods to achieve consistency in
design. It is crucial that the vent system for a reactor should
be designed for the worst case scenario which is reason-
ably foreseeable. The scenarios considered should include fire
engulfment as well as reaction runaway (Mannan, 2012).
3.7. Process control and quenching
Quenching is a method of inhibiting runaway reactions. A cold
inert diluent is inserted into the reaction mixture that stops
the reaction by cooling it. The ideal substance is water, assum-
ing that it could be applicable to particular system as it is
cheap, readily available and has a high specific heat capacity.
However, in some processes water may react exothermically
with the reaction mixture and therefore alternative diluents
should be used.
In an industry, the most common method of addition of
diluent is by quenching, where the diluent is added rapidly
to the reactor from a storage vessel mounted above it. When
a runaway reaction is detected in the reactor (indicated by a
temperature indicator), a valve opens automatically and the
quench liquid runs rapidly into the reactor under gravity. The
quench valve should be opened and should not fail during
runaway reactions. In some cases, where adequate quantity
of diluent could not be introduced into the reactor because
of less space within the reactor, dumping should be used
(Prugh and Johnson, 1988). The advantages of quenching or
dumping include no environmental pollution since nothing is
discharged to the environment, the quenched mixture could
be easily disposed off and the recovery of expensive reactants
is possible.
3.8. Process control and containment
Containment of the products of a runaway reaction has
been an attractive option that prevents toxic, flammable and
corrosive decomposition products from escaping to the envi-
ronment causing environmental hazards. A reactor vessel for
total containment should be designed to resist the maximum
pressure produced during the runaway reaction. Particular
consideration has to be given to weaker parts of the reactor
system, e.g., associated plant item such as a glass condenser.
3.9. Assessment strategy
Runaway reactions occur in a batch reactor in an industry
due to the loss of thermal control. A thermal hazards assess-
ment strategy has been used to identify the materials and
unit processes that have been potentially hazardous and to
quantify the hazards, which arise from these assessments
with a minimum of testing (Callaghan, 2001). A thermal haz-
ard assessment involves a sequential approach that covers
thermochemical evaluation, reaction calorimetry (Regestein
et al., 2012), effects caused by scale-up, accumulation, cooling
failure, heat flow calculations including normal reaction
data (H for the reaction) and in the worst case scenario for
possible thermal runaways.
Process risk management strategies could be used in
order to reduce risks, frequency or consequences of potential
accidents. Process risks could be managed using either inher-
ent or passive strategies.
4. Conclusions
A critical analysis of thermal runaway reactions in the UK
has been carried out for the last 25 years (1988–2013) and
the results have been compared with the earlier publications.
The findings from the current research clearly indicated that
lessons have not been learnt as a consequence of runaway
reactions. Even though the number of incidents has been
reduced in the last 25 years, the consequence of runaway
reactions has become more serious leading to an increased
number of fatalities and injuries. These revelations are in
complete contrast to the amount of money and resources all
industries have been spending in terms of gaining advanced
knowledge in technical knowhow of operating a chemical
plant on a day-to-day basis to avoid incidents caused by
thermal runaway. It is evident that teaching in the aca-
demic institutions as well as training in the industries lack
in polymerisation and decomposition reactions. This study
highlights that similar trends have been observed over the
past five decades for thermal runaway incidents caused by
mischarging, agitation and quality control. This study also
indicates that operator errors, management failures and the
absence of organised procedures have been the major con-
tributors for runaway incidents. It is recommended that all
industries should introduce systematic assessment proce-
dures for chemical reaction hazard and follow it in order to
assist in reducing the most common types of errors attributed
to thermal runaway reactions. It is essential that operators
have a thorough training including a thorough understand-
ing of the process chemistry and thermochemistry, ensuring
adequate engineering design for heat transfer, adequate con-
trol systems (temperature control), safety back-up systems
and adequate operational procedures. Managers should not
neglect the safety by reducing the number of operating staff,
thereby increasing the likelihood of an incident, to generate
more profit. In fact, managers should discuss and consider
the probability and consequences of an increase in the risk by
reducing the number of operating staff. It is also advised that
managers should be aware that safe operation and success of
sustainable business could not be separated.
Acknowledgements
Rim Saada and Dipesh Patel gratefully acknowledge the finan-
cial support provided by London South Bank University, UK.
The authors acknowledge the support and guidance provided
by the late Professor Philip Nolan of London South Bank Uni-
versity.
References
Ahmad, W., Baloch, M.K., 2007. The effect of inactive impurities
on a surface in NO–CO reaction: a Monte Carlo simulation.
Appl. Surf. Sci. 253, 8447–8452.
Barton J.A. and Nolan P.F, 1991, Incidents in the chemical
industries due to thermal runaways chemical reactions, In
‘Safety of chemical batch reactors and storage tanks’, Eds.
Benuzzi A. and Zaldívar, J.M., Kluver Academic Publishers,
Dordrecht, The Netherlands, 99-124.
Barton, J.A., Rogers, R., 1997. Chemical Reaction Hazards, second
ed. Institution of Chemical Engineers, UK, 2, 13, 123.
 Process Safety and Environmental Protection 9 7 ( 2 0 1 5 ) 109–115
Callaghan, B., 2001. Reducing Risks, Protecting People. HSE, UK,
Available from http://www.hse.gov.uk/risk/theory/
r2p2.pdf (accessed 12.08.2014).
FACTS, 2006. Accident Number: 23230. Chemical Accident in
United Kingdom with Sodium Dichlorocyanurate. FACTS,
Available from http://www.factsonline.nl/accident/23230/
chemical-accident-in-united-kingdom-with-sodium-
dichlorocyanurate (accessed
12.08.2014).
FACTS, 1999. Accident Number: 17393. Chemical Accident in
United Kingdom with 2-cl-5-cl-Methyl Thiazole/ccmz. FACTS,
Available from http://www.factsonline.nl/accident/17393/
chemical-accident-in-united-kingdom-with-2-cl-5-cl-methyl-
thiazole-ccmz (accessed
12.08.2014).
FACTS, 1994. Accident Number: 18732. Chemical Accident in
United Kingdom with Hydrogen (gas). FACTS, Available from
http://www.factsonline.nl/accident/18732/chemical-
accident-in-united-kingdom-with-hydrogen-(gas) (accessed
12.08.2014).
FACTS, 1992a. Accident Number: 12060. Chemical Accident in
United Kingdom with Sludge Residues. FACTS, Available from
http://www.factsonline.nl/accident/12060/chemical-
accident-in-united-kingdom-with-sludge-residues (accessed
12.08.2014).
FACTS, 1992b. Accident Number: 13385. Chemical Accident in
United Kingdom with Carbon Tetrachloride. FACTS, Available
from http://www.factsonline.nl/accident/13385/chemical-
accident-in-united-kingdom-with-carbon-tetrachloride
(accessed 12.08.2014).
FACTS, 1992c. Accident Number: 15501. Chemical Accident in
United Kingdom with Ethene (pressurized). FACTS, Available
from http://www.factsonline.nl/accident/15501/chemical-
accident-in-united-kingdom-with-ethene-(pressurized)
(accessed on 12.08.2014).
FACTS, 1991. Accident Number: 11381. Chemical Accident in
United Kingdom with Toluene (Toluol/Methylbenzene).
FACTS, Available from http://www.factsonline.nl/accident/
11381/chemical-accident-in-united-kingdom-with-toluene-
(toluol-methylbenzene) (accessed
12.08.2014).
FACTS, 1990. Accident Number: 10330. Chemical Accident in
United Kingdom with Dimethyl Acetamide (dmac). FACTS,
Available from http://www.factsonline.nl/accident/10330/
chemical-accident-in-united-kingdom-with-dimethyl-
acetamide-(dmac) (accessed
12.08.2014).
FACTS, 1989. Accident Number: 11639. Chemical Accident in
United Kingdom with Chlorine. FACTS, Available from
http://www.factsonline.nl/accident/11639/chemical-
accident-in-united-kingdom-with-chlorine (accessed
12.08.2014).
FACTS, 1988. Accident Number: 13457. Chemical Accident in
United Kingdom with ortho-Nitrobenzylnitrate. FACTS,
Available from http://www.factsonline.nl/accident/13457/
chemical-accident-in-united-kingdom-with-ortho-
nitrobenzylnitrate (accessed
12.08.2014).
Fishwick, T., 2012. Lessons reLearned-Recurring Accidents:
Confined Spaces. tcetoday, Rugby, UK, pp. 30–34.
Fishwick, T., 2008. A runaway reaction results in multiple
fatalities and injuries. Loss Prev. Bull. 201, 10–12.
Fishwick, T., 2004. An uncontrolled chemical decomposition. Loss
Prev. Bull. 177, 8–9.
115
Gakhar, S., Carr, D., 2003. Alkoxylation runaway reaction incident
at Baker Petrolite, Hartlepool, UK. Loss Prev. Bull. 174, 3–10.
Gillard, T., 1998. Loss of reactor contents to atmosphere. Loss
Prev. Bull. 143, 21–22.
Gustin, J., 2002. Influence of trace impurities on chemical
reaction hazards. J. Loss Prev. Process. Ind. 15 (1), 37–48.
Hellemans, M., 2009. The Safety Relief Valve Handbook: Design
and Use of Process Safety Valves to ASME and International
Codes. Elsevier Inc., UK, pp. 18.
HSE, 1998. Chemical Reaction Hazards and the Risk of Thermal
Runaway. HSE, UK, Available from
http://www.hse.gov.uk/pubns/indg254.htm (accessed
12.08.2014).
Kletz, T., 1995. Some accidents of 2005. Loss Prev. Bull. 120, 5–7.
Kletz, T., 1991. The 1990 Shell explosion. Loss Prev. Bull. 100,
21–25.
Lindley, J., 2001. Explosion of an ethoxylation reactor. Loss Prev.
Bull. 157, 19–20.
LPB, 1996. High pressure reactor rupture. Loss Prev. Bull. 128, 3–5.
LPB, 1995. Acrylic acid runaway. Loss Prev. Bull. 126, 3–6.
LPB, 1994a. Explosion in a nitration reaction vessel. Loss Prev.
Bull. 115, 1–5.
LPB, 1994b. Runaway polymerisation leads to release of vinyl
acetate. Loss Prev. Bull. 118, 8.
LPB, 1993a. Runaway cleanout. Loss Prev. Bull. 109, 7–10.
LPB, 1993b. Polymerisation runaway reaction. Loss Prev. Bull. 110,
25–26.
LPB, 1993c. Emulsion polymerisation incident. Loss Prev. Bull.
113, 25.
LPB, 1992. Fluid catalytic cracker regenerator incident. Loss Prev.
Bull. 107, 16–20.
LPB, 1990. Incompatible material causes explosion in waste
container. Loss Prev. Bull. 91, 15.
Mannan, S., 2012. Lees’ Loss Prevention in the Process Industries,
fourth ed. Elsevier Inc., UK, 33, 469, 507, 1467.
Pinsky, M.L., Vickery, T.P., Freeman, K.P., 1990. Investigation of a
drum explosion of 30% aqueous KOCN. J. Loss Prev. Process
Ind. 3, 345–348.
Prugh, R.W., Johnson, R.W., 1988. Guidelines for Vapour Release
Mitigation. American Institute of Chemical Engineers, USA,
pp. 48.
Regestein, L., Heiner, G., Zavarel, M., Buchs, J., 2012. Comparison
of two methods for designing calorimeters using stirred tank
reactors. Biotechnol. Bioeng. 110 (1), 180–190.
Sachdev, A., Todd, J., 2005. Incident investigation of mono-nitro
toluene still explosion. J. Loss Prev. Process. Ind. 18 (4),
531–536.
Shu, C.M., Wang, C.J., 2003. An incident investigation on an
unexpected loss of containment and damage to PHI-TEC II
during acrylonitrile runaway tests. J. Loss Prev. Process Ind. 16
(6), 593–595.
Travers, I., Smith, N., 2012. Process Safety-Educate Your Boss!.
tcetoday, Rugby, UK, pp. 24–26.
Vasanthi, D., Pranavamoorthy, B., Pappa, N., 2012. Design of a
self-tuning regulator for temperature control of a
polymerization reactor. ISA Trans. 51, 22–29.
Van Reijendam, J.W., Verbrugge, H., Ten Brink, J.V., Coolegem, J.,
Heijmen, H.J., Winter, M., 1992. Explosion in the Shell
fluoroaromatics plant at Stanlow, March 1990: an
investigation of the underlying chemistry. J. Loss Prev. Process
Ind. 5 (4), 211–213.
Whitmore, M.W., Gladwell, J.P., Rutledge, P.V., 1993. Report of an
explosion during the manufacture of an azodicarbonamide
formulation. J. Loss Prev. Process Ind. 6 (3), 169–175.