Journal of Loss Prevention in the Process Industries 15 (2002) 291303

www.elsevier.com/locate/jlp

Review of 62 risk analysis methodologies of industrial plants

J. Tixier a,, G. Dusserre a, O. Salvi b, D. Gaston b
a

Ecole des Mines dAle`s, Laboratoire Genie de lEnvironnement Industriel, 6 Avenue de Clavie`res, 30319 Ale`s Cedex, France
b
INERIS-Direction des Risques Accidentels, Parc Technologique ALATA B.P. no. 2, 60550 Verneuil-en-Halatte, France

Abstract
For about 10 years, many methodologies have been developed to undertake a risk analysis on an industrial plant. In this paper,
62 methodologies have been identified, these are separated into three different phases (identification, evaluation and hierarchisation).
In order to understand their running, it seems necessary to examine the input data, methods used, obtained output data and to rank
them in several classes. First, all the input data are grouped together into seven classes (plan or diagram, process and reaction,
products, probability and frequency, policy, environment, text, and historical knowledge). Then, the methods are ranked in six
classes based on the combination of four usual criteria (qualitative, quantitative, deterministic and probabilistic). And finally, the
output data are classified into four classes (management, list, probabilistic and hierarchisation). This classification permits the
appraisal of risk analysis methodologies. With the intention of understanding the running of these methodologies, the connections
between the three defined previously criteria (determinist, probabilistic and determinist and probabilistic) are brought to the fore.
Then the paper deals with the application fields and the main limitations of these methodologies. So the hierarchisation phase is
discussed and the type of scale used. This paper highlights the difficulties in taking into account all risks for an industrial plant
and suggests that there is not only one general method to deal with the problems of industrial risks. 2002 Elsevier Science Ltd.
All rights reserved.
Keywords: Industrial hazards; Risk assessment; Explosions; Fires; Toxic gas dispersion; Hierarchisation

1. Introduction
The industrial risk problem and the diversification of
risk types have increased concurrently with industrial
development. In the same time, the risk acceptability
threshold of the population has decreased. In response
to this preoccupation, competent authorities and industrialists have developed methodologies and tools for risk
prevention and protection, as well as crisis management.
To cope up with major accidents, a previous analyses
should be done. The forward-looking risk analysis permits an exhaustive identification of potential hazardous
sources to prevent accident scenarios and to assess
potential impact on human, environmental and equipment targets in order to propose prevention or protection
(Lagadec, 1980). The risk analysis methodologies
focuses on the main hazard sources. Two principal
sources of risk can be brought to the fore: industrial

establishment and transport of dangerous goods. These

two types of sources are quite different. At first sight,
the quantities involved are not really comparable, and
the environment is unsettled for an industrial site whilst
the opposite is true for the case of transport of dangerous goods.
So to analyse and to manage safety aspects, various
approaches are proposed, they focus on organisational
and technical features. Sixty-two risk analysis methodologies are set out in the following.
2. Risk analysis methodologies
The management of major industrial risk should be
one of the most important preoccupations for operators.
To deal with this problem, many risk analysis methodologies were developed by industrialists and competent
authorities.
2.1. Description of these methodologies

Corresponding author. Tel.: +33-4-66-78-27-53; fax: +33-4-6678-27-01.

E-mail address: jerome.tixier@ema.fr (J. Tixier).

This work identified more than 60 risk analysis methodologies, which can include up to three main phases:

0950-4230/02/$ - see front matter 2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 9 5 0 - 4 2 3 0 ( 0 2 ) 0 0 0 0 8 - 6

292

J. Tixier et al. / Journal of Loss Prevention in the Process Industries 15 (2002) 291303

An identification phase based on a site description

(hazardous activities, products and equipment). Those
data are necessary to develop the processes of the
methodologies.
An evaluation phase to realise a quantification of the
risk. There are two ways to lead thisa deterministic
approach and/or a probabilistic approach. This evaluation gives the previously found consequences of
scenarios and enables their impacts on the industrial
site or on its vicinity to be taken into account.
A hierarchisation phase which aims at ranking some
results, obtained through the two previous phases, in
order to put preponderant risks forward. Thanks to
this hierarchisation, the most important risks could be
solved first.
The phase of risk identification is essential, because
it establishes the bases of the risk analysis. Indeed, the
data of risk identification will be the input of the evaluation and/or hierarchisation phases. Therefore, it is
necessary to make an identification phase in an exhaustive way to get the best results.
The phase of risk evaluation should be realised
according to two different approaches: either by the
evaluation of damage consequences (deterministic
approach) or by the evaluation of accident probability
(probabilistic approach). The phase of risk hierarchisation ranks the risks obtained in the previous phase, in
order to implement modifications or corrective actions
on the most severe risk systems.
A risk analysis methodology does not necessarily contain these three phases. It can be constituted by only the
following combinations: an identification phase; identification and evaluation phases; or identification, evaluation and hierarchisation phases. Whatever methodologies are used, to carry out a risk analysis, three kinds
of elements are required: the expected output data; available input data; and the selected method. Indeed, users
propose some objectives to reach (output data expected),
next they collect information concerning the studied system (input data available), and finally they choose the
applied method according to the two previous elements.
2.1.1. Types of method
The used methods can be described according to the
following four properties:

Deterministic
Probabilistic
Qualitative
Quantitative

These methodologies can be sorted out in two principal

groups, one qualitative and the other quantitative. Each
group can be divided into three categories: only deter-

ministic; only probabilistic; and a combination of deterministic and probabilistic approach.

The deterministic methods take into consideration the
products, the equipment and the quantification of consequences for various targets such as people, environment
and equipment. The probabilistic methods are based on
the probability or frequency of hazardous situation
apparitions or on the occurrence of potential accident.
The probabilistic methods are mainly focused on failure
probability of equipment or their components. On the
one hand, probabilistic methods are used to lead an
analysis on a restricted part of a plant. On the other hand,
deterministic and combined deterministic and probabilistic methods are used to analyse the whole industrial
establishment.
The classification of the methods is based on the type
of output data. In each category, methods can be ranked
from the simple, which comprises only one step to the
more complex ones that are based on the three steps
(identification, evaluation and hierarchisation phases).
The complex methods are generally composed of modules issued from simple methods and other modules are
added in order to realise a more complete risk analysis
with easier results to analyse.
In Table 1, 62 methodologies are ranked according to
the four defined previously criteria. The great majority of
methods are deterministic, because historically operators
and public organisations have initially tried to quantify
damages and consequences of potential accidents, before
to understand why and how they could occur.
2.1.2. Types of input data
Input data can be technical like process characteristics
or qualitative like safety policy. The analysis of 62
methods leads us to propose seven classes of input data.
For each class some information is given on the type of
input data.
Plans or diagrams are related to the description of
the site, installation, units, fluid networks, safety barriers and the storages.
Process and reactions are related to operations and
tasks description, physical and chemical features of
process, kinetic and calorimetric parameters,
operating conditions and normal functioning conditions.
Substances are related to the type of substance, physical and chemical properties, quantities and the toxicological data.
Probability and frequency are related to the type of
failure, probability and frequency of failure, human
failure, failure rate and the exposure probability.
Policy and Management are related to the maintenance, organisation, safety policy, Safety Management System, transport management and the equipment cost.

J. Tixier et al. / Journal of Loss Prevention in the Process Industries 15 (2002) 291303

293

Table 1
Classification of risk analysis methodologies
Risk analysis methodologies

Deterministic

No.a Qualitative

No. Quantitative

Action Errors Analysis AEA (Rogers, 2000)

31

Checklist Khan & Abbasi, 1998b

32

Concept Hazard Analysis CHA (Rasmussen & 33

Whetton, 1997; Rogers, 2000)
Concept Safety Review CSR (Rogers, 2000)
34

4
5

6
7

8
9
10
11

12
13
14
15

16
17

18
19
20
21
22

Failure Mode Effect Analysis FMEA (Khan &

Abbasi, 1998b; Nicolet-Monnier, 1996;
Rogers, 2000)
Goal Orinted Failure Analysis GOFA (Rogers,
2000)
Hazard and Operability HAZOP (Kennedy &
Kirwan, 1998; Khan & Abbasi, 1998b;
Nicolet-Monnier, 1996; Rogers, 2000;
Tweeddale, Cameron, & Sylvester, 1992)
Human Hazard and Operability HumanHAZOP
(Kennedy & Kirwan, 1998)
Insurers involvement in risk reduction process
(Sankey, 1998)
Manager (Pitblado, Williams, & Slater, 1990)

35

36
37

38
39
40

Optimal Hazard and Operability OptHAZOP

41
(Khan & Abbasi, 1997a; Khan & Abbasi,
1998b)
Plant Level Safety Analysis PLSA (Toola,
42
1992)
Potential domino effects identification
43
(Delvosalle, Fievez, & Benjelloun, 1998)
Preliminary Risks Analysis PRA (NicoletMonnier, 1996; Rogers, 2000;)
Process Risk Management Audit PRIMA
Hurst, Young, Donald, Gibson, & Muyselaar,
1996
Profile Deviation Analysis PDA (Korjusiommi,
Salo, &Taylor, 1998)
Safety related questions for computer
controlled plants (Chung, Broomfield, & Yang,
1998; Yang & Chung, 1998)
Seqhaz Hazard Mapping SHM (Korjusiommi
et al., 1998)
Sneak Analysis (Rogers, 2000)
Task Analysis TA (Rogers, 2000)
What if? Analysis (Khan & Abbasi, 1998b;
Nicolet-Monnier, 1996; Rogers, 2000)
World Health Organisation WHO (Khan &
Abbasi, 1998b)

Accident Hazard Analysis AHI (Khan & Abbasi,

1997b; Khan & Abbasi, 1998a)
Annex 6 of SEVESO II Directive (La directive
Seveso II: Annexe 6, 1997)]
Chemical Runaway Reaction Hazard Index RRHI
(Kao & Duh, 1998)
Dows Chemical Exposure Index CEI (American
Institute of Chemical Engineers, 1994)
Dow Fire and Explosion Index FEI (American
Institute of Chemical Engineers, 1987; Khan &
Abbasi, 1998a)
Fire and Explosion Damage Index FEDI (Khan &
Abbasi, 1998a)
Hazard Identification and Ranking HIRA (Khan &
Abbasi, 1997b; Khan & Abbasi, 1998b)

Instantaneous fractionnal loss index IFAL

(Khan & Abbasi, 1998a; Khan & Abbasi, 1998b)
Methodology of domino effects analysis
(Dolladille, 1999)]
Methods of potential risk determination and
evaluation (Ja ger & Ku hnreich, 1998)
Mond Fire Explosion and Toxicity Index FETI
(Khan & Abbasi, 1998a; Khan & Abbasi, 1998b)
SAATY methodology (Troutt & Elsaid, 1996)
Toxic Damage Index TDI (Khan & Abbasi,
1998a)

(continued on next page)

Environment is related to the site environment, topographical data and the population density.
Text and historical knowledge are related to the standards and regulations, and historical knowledge.

Most of the methods are based on a general description of the site (Plans and Diagrams) and a few take
into account the Environment.The common input data
such as:

These input data are summarised in Table 2. The latter

gives connections between input data and the listed
methodologies.

plan or diagram;
process and reactions;
products

294

J. Tixier et al. / Journal of Loss Prevention in the Process Industries 15 (2002) 291303

Table 1 (continued)
Risk analysis methodologies
No.a Qualitative
Probabilistic

23
24

25

No. Quantitative

Accident Sequences Precursor ASP (Holmberg, 44

1996)
Delphi Technique (Rogers, 2000)
45

Earthquake safety of structures and

installations in chemical industries (Jezler,
1998)

46

47
48
49
Deterministic and
probabilistic

26
27
28
29
30

Maximum Credible Accident Analysis MCAA 50

(Khan & Abbasi, 1998b)
Reliability Block Diagram RBD (Rogers,
51
2000)
Safety Analysis SA (Khan & Abbasi, 1998b) 52
Safety Culture Hazard and Operability
SCHAZOP (Kennedy & Kirwan, 1998)
Structural Reliability Analysis SRA (Rogers,
2000)

53
54
55
56
57
58

59

60
61
62
a

Defi method (Rogers, 2000)

Event Tree Analysis ETA (Gadd, Leeming, &
Riley, 1998; Nicolet-Monnier, 1996; Rogers, 2000;
Tiemessen & van Zweeden, 1998;)
Fault Tree Analysis FTA (Khan & Abbasi, 1998b;
Nicolet-Monnier, 1996; Rogers, 2000)
Maintenance Analysis MA (Rogers, 2000)
Short Cut Risk Assessment SCRA (Rogers, 2000)
Work Process Analysis Model WPAM
(Davoudian, Wu, & Apostolakis, 1994)
AVRIM2 (Ham, van Kessel ,& Wiersma, 1998)
Facility Risk Review (Schlechter, 1996)
Failure Mode Effect Criticality Analysis FMECA
(Rogers, 2000)
IDEF3 (Kusiak & Zakarian, 1996; Larson &
Kusiak, 1996)
International Study Group on Risk Analysis
ISGRA (Khan & Abbasi, 1998b)
IPO Risico Berekening Methodiek IPORBM
(Tiemessen & van Zweeden, 1998)
Method Organised Systematic Analysis of Risk
MOSAR (Perhillon, 2000; Rogers, 2000)
Optimal Risk Assessment ORA (Khan & Abbasi,
1998b)
Probabilistic Safety Analysis PSA (Khan &
Abbasi, 1998b; Papazoglou, Noivolianitou,
Aneziris, & Christou, 1992)
Quantitative Risk Assessment QRA (Khan &
Abbasi, 1998b; Leeming & Saccomanno, 1994;
Oien, Sklet, & Nielsen, 1998; Puertas, Sanz,
Vaquero, Marono, & Sola, 1998; Rogers, 2000)
Rapid Ranking RR (Larson & Kusiak,
1996Tweeddale et al., 1992)
Rapid Risk Analysis Based Design RRABD
(Khan & Abbasi, 1998)
Risk Level Indicators RLI (Oien et al., 1998)

Each methodology is referred to by a number.

are generally used for deterministic methods, while

deterministic and probabilistic methods use as mainly
input data:
probability and frequency;
plan or diagram;
products.
The more specific input data from Table 2 (Policy and
Management, environment, and texts and historical
knowledge) are principally employed in deterministic
methods.

2.1.3. Types of output data

Output data can be qualitative like recommendations
or quantitative like index of risk level. From the review
of 62 methods, four classes of output data proposed are
as follows.
Management is related to actions, recommendations,
modifications, and formation or operation procedures.
Lists are related to lists of errors, hazards, domino
effects, causes/consequences, failures and damages,
critical activities, failure mode, accident initiators,
vulnerable place and major accident scenarios.

J. Tixier et al. / Journal of Loss Prevention in the Process Industries 15 (2002) 291303

295

Table 2
Connections between input data and methodologies
Types

Input data

Methodologiesa
Deteministic

Probabilistic

Det and prob

Site

Qualitative

2, 3, 4, 5, 7, 8, 11, 13,
14, 15, 18, 19, 21, 22

Installations
Units
Fluid or gas networks
Functionning

Quantitative

31, 34, 35, 36, 37, 38, 39, 45, 46, 47

40, 41, 42, 43

Plan and diagram

23

26, 27, 30

50, 51, 52, 53,

56, 57, 59, 61, 62

Safety barriers
Storages
Process and reactions
Operations description
Qualitative
Tasks description
Reactions and physical
and chemical features
Process characteristics
Kinetics and
Quantitative
calorimetrical parameters
Normal functionning
conditions
Operating conditions

2, 3, 7, 8, 10, 11, 12, 14, 16, 19, 20, 21

33, 35, 36, 37, 40, 41, 42, 43

54

11, 12, 13, 14, 16,

26

31, 32, 33, 34, 35, 36, 38, 40, 41, 42, 43

51, 54, 55, 56, 58

1, 12, 23, 24

26, 27, 28, 29, 30

Products
Products types, physical qualitative
and chemical properties
Quantities
toxicological data
Quantitative
Probability and frequency
Failure type
Failure probability
Initiation and failure
frequencies
human failure

Qualitative

Quantitative

25

44, 46 ,47, 48, 49 50, 51, 52, 53,

58, 59, 60

Failure rate
Exposure probability
Policy and Management
Maintenance
Organisation
Safety policy
SMS
Transport Management
Equipments cost

Qualitative

2, 9, 10, 15, 17

29

Quantitative

36, 42

Qualitative

11, 19

Quantitative

31, 34, 36, 37, 42

51

Qualitative
Quantitative

4, 5, 13, 14, 18
31, 35, 39, 41, 42

51, 52, 59, 61

49

51, 55

Environment
Site environment
Topographical data
population density
Texts and historical knowledge
standards
Regulations and
Documents
Historical knowledge
a

The number refers to the methodologies presented in Table 1.

296

J. Tixier et al. / Journal of Loss Prevention in the Process Industries 15 (2002) 291303

Probabilistic is related to failure rate, reliability, scenarios or damages probability, and accident frequency.
Hierarchisation is related to level risk index, severity
and criticality, fire, explosion, toxic leakage index,
organisational index, classification according to the
type of risk.

2.1.4. Links between input data, output data and

methods
Now, it is relevant to underline how links between
input data, method and output data are running. Table 4
can be used according to whether the user expects some
results or has some available data:

The connections between output data and methodologies are presented in Table 3. Output data like Management and Lists are based on expert choices and give
qualitative results while output data like Probabilistic
and Hierarchisation give quantitative results. The results
are mainly in the form of lists. The deterministic and
probabilistic methods provide several types of results, so
more information can be obtained with this approach.

First, if industrialists need a certain type of results,

then they will read through the results (output data)
columns given in Table 4. So different types of
methods are proposed and finally the necessary input
data can be identified
Secondly, if only several input data are available, then
the user will read through the input data columns
given in Table 4. The combination of available input

Table 3
Links between output data and methodologies
Types

Methodologiesa

Output data

Deteministic

Probabilistic

Det and prob

Management
Actions
Recommendations
modifications
Formation and
operation procedures

Qualitative

3, 4, 5, 6, 7, 9, 10, 11,
15, 16, 20

Quantitative

39

50, 51, 59, 60

Qualitative

1, 2, 5, 6, 7, 8, 11, 12, 23, 24

13, 14, 17, 19, 21, 22

26, 28, 29, 30

39

50, 51, 53, 54, 56, 57,

59, 60, 61

List
List of errors

Estimation/list of risks
List of domino effects
List cause/consequence
failure, damage
List of installation
critical activities
List of failure mode
Quantitative

46, 47, 48, 49

List accident initiators

List of vulnerable
place
List of major scenarios
Probabilistic
Failure rate
Reliability
Scenarios or damages
probability
Accident frequency

Qualitative

23, 24, 25

b
26, 27, 28, 29, 30

Quantitative

44, 45, 49

53, 54

Hierarchisation
Risk index/level
Severity/criticity
Fire/explosion index
Toxic leakage index

Qualitative

11, 18

Quantitative

31, 32, 33, 34, 35, 36,

37, 38, 40, 41, 42, 43

Organisational risk
index
Type risk classification
a

The number refers to the methodologies presented in Table 1.

25

51, 52, 55, 56, 58, 61,

62

The number refers to the methodologies presented in Table 1.

Table 4
Links between input data, methods and output dataa

J. Tixier et al. / Journal of Loss Prevention in the Process Industries 15 (2002) 291303
297

298

J. Tixier et al. / Journal of Loss Prevention in the Process Industries 15 (2002) 291303

data permits the identification of methods which are

conceivable to use in the risk analysis.
Table 4 is a synthesis of this study and a tool for an
identification of methods which could be used according
to objectives and available input data.The analysis of
Table 4 highlights that many input data are necessary to
realise a:
qualitative and deterministic risk analysis;
quantitative and deterministic risk analysis;
quantitative and deterministic and probabilistic risk
analysis.
Whatever qualitative or quantitative methods, results are
complete when both deterministic and probabilistic
methods are used.
Probabilistic methods need some input data, but they
do not take into account some specificities of the industrial establishment like Policy or Environment. Now, the
running of methodologies have been brought to the fore,
and it is necessary to discuss two important points: on
the one hand, the application fields of those methodologies and on the other hand, their main limitations.
2.2. Application fields of methodologies
The application field of these different methodologies
can be ranked into three categories (Table 5). First,
which is the most important in number of developed
methodologies, concerns industrial site. Generally, some
methodologies are developed for specific application or
process and they are not transposable to different types
of industrial establishment. The second application field
is the transportation of dangerous goods and the third
one permits to take into account human factors in a specific environment. Some methodologies can be used for
various purposes and several application fields. For
example among the most general methods, there are
What if and Safety Analysis.
2.3. Limitations of methodologies
The main limitations of those methodologies can be
summarised in the following points.
The more general the methodology is, the less it takes
into account the specificities of the studied case.
On the contrary, if the methodology is too specific it
will be less transposable to another case.
Knowledge of people, who are participating in the
risk analysis, is quite important (different types of
competences and levels of people involvement).
For probabilistic analysis, the validity of data is a
decisive parameter.
The updating of data takes a lot of time work.

For some methodologies, the operational application

is difficult to realise because of the lack of description. It is useful to provide a guide book to explain
how methodologies could be used.
The complexity of methods requires specific training
for their implementation.
It can be noticed that there is a great disconnection
between risk analysis methodologies and human factors.
A new trend highlights a need to have more quickly
exploitable results. So, the output data of evaluation
phase must be completed by an additional step, i.e. the
hierarchisation phase.

3. The risk hierarchisation

The recent evolution in risk analysis methodologies
shows that easily applicable methods are proposed with
a risk level index as a result. The hierarchisation consists
of the action to organise some elements, data, or events
in increasing (or decreasing) order with the help of
classification tabs, with the view to bringing out the main
points for analysis. These methodologies are simple and
rapid to use, the specific hierarchisation rules in the
shape of data table are usually provided by the authors
methods for the hierarchisation phase. The association
between risk level and geographical zones on site is primordial to have a good safety management.
These different methodologies including hierarchisation step are gathered in Table 6. Methodologies with
hierarchisation phase are almost all quantitative and
deterministic. They are based on the development of a
risk level index, calculated for each element, unit or
studied area in order to rank them. This ranking provides
the way to clearly identify critical areas on industrial
site. Thus it permits to realise priority actions with the
view to decreasing the probability of occurrence
(prevention) or to reducing the consequences of an accident (protection including emergency response).
To calculate this risk level index, several parameters
characterising the site and processes are identified. Then
these parameters are ranked with the help of scales
which are provided in a guide and which can be based
on a deterministic or probabilistic approach.
The deterministic scales can be quantitative or qualitative for the following variables.
Internal hazard of substances and equipments:
reaction types (hydrolysis, oxidation, reduction,
polymerisation,);
reaction parameters (stability, reactivity, exothermically, pressure or temperature of reaction,);
physical and chemical properties of substances, their

Industrial site

Method Organised Systematic Analysis of Risk MOSAR (Perhillon, 2000; Rogers,

2000)
Methodology of domino effects analysis (Dolladille, 1999)]
Methods of potential risk determination and evaluation (Ja ger & Ku hnreich, 1998)
Mond Fire Explosion and Toxicity Index FETI (Khan & Abbasi, 1998a; Khan &
Abbasi, 1998b)
Optimal Hazard and Operability OptHAZOP (Khan & Abbasi, 1997a; Khan &
Abbasi, 1998b)
Optimal Risk Assessment ORA (Khan & Abbasi, 1998b)
Plant Level Safety Analysis PLSA (Toola, 1992)
Potential domino effects identification (Delvosalle et al., 1998)
Preliminary Risks Analysis PRA (Nicolet-Monnier, 1996; Rogers, 2000]
Probabilistic Safety Analysis PSA (Khan & Abbasi, 1998b; Papazoglou et al., 1992)
Profile Deviation Analysis PDA Korjusiommi et al., 1998

Quantitative Risk Assessment QRA (Alonso & Gavalda, 1998; Khan & Abbasi,
1998b; Leeming & Saccomanno, 1994; Oien et al., 1998; Papazoglou et al., 1992;
Puertas et al., 1998; Rogers, 2000)
Earthquake safety of structures and installations in chemical industries (Jezler, 1998) Rapid Ranking RR (Larson & Kusiak, 1996; Tweeddale et al., 1992)
Event Tree Analysis ETA (Gadd et al., 1998; Nicolet-Monnier, 1996; Rogers, 2000; Rapid Risk Analysis Based Design RRABD (Khan & Abbasi, 1998)
Tiemessen & van Zweeden, 1998)
Facility Risk Review (Schlechter, 1996)
Reliability Block Diagram RBD (Rogers, 2000)
Failure Mode Effect Analysis FMEA (Khan & Abbasi, 1998b; Nicolet-Monnier,
Risk Level Indicators RLI (Oien et al., 1998)
1996; Rogers, 2000)
Failure Mode Effect Criticality Analysis FMECA (Rogers, 2000)
SAATY methodology (Troutt & Elsaid, 1996)
Fault Tree Analysis FTA (Khan & Abbasi, 1998b;Nicolet-Monnier, 1996; Rogers,
Safety Analysis SA (Khan & Abbasi, 1998b)
2000)
Fire and Explsion Damage Index FEDI (Khan & Abbasi, 1998a)
Safety Culture Hazard and Operability SCHAZOP (Kennedy & Kirwan, 1998)
Goal Orinted Failure Analysis GOFA (Rogers, 2000)
Safety related questions for computer controlled plants (Chung et al., 1998; Yang &
Chung, 1998]
Hazard and Operability HAZOP (Kennedy & Kirwan, 1998; Khan & Abbasi, 1998b; Seqhaz Hazard Mapping SHM Korjusiommi et al., 1998
Nicolet-Monnier, 1996; Rogers, 2000; Tweeddale et al., 1992;)
Hazard Identification and Ranking HIRA (Khan & Abbasi, 1997b; Khan & Abbasi, Short Cut Risk Assessment SCRA (Rogers, 2000)
1998b)
IDEF3 (Kusiak & Zakarian, 1996; Larson & Kusiak, 1996)
Sneak Analysis (Rogers, 2000)
Instantaneous fractionnal loss index IFAL (Khan & Abbasi, 1998a; Khan & Abbasi, Structural Reliability Analysis SRA (Rogers, 2000)
1998b)
Insurers involvement in risk reduction process (Sankey, 1998)
Toxic Damage Index TDI (Khan & Abbasi, 1998a)
International Study Group on Risk Analysis ISGRA (Khan & Abbasi, 1998b)
What if? Analysis (Khan & Abbasi, 1998b; Nicolet-Monnier, 1996; Rogers, 2000)
Maintenance Analysis MA (Rogers, 2000)
World Health Organisation WHO (Khan & Abbasi, 1998b)
Maximum Credible Accident Analysis MCAA (Khan & Abbasi, 1998b)
(continued on next page)

Chemical Runaway Reaction Hazard Index RRHI (Kao & Duh, 1998)
Concept Hazard Analysis CHA (Rasmussen & Whetton, 1997; Rogers, 2000)
Concept Safety Review CSR (Rogers, 2000)
Defi method (Rogers, 2000)
Delphi Technique (Rogers, 2000)
Dows Chemical Exposure Index CEI (American Institute of Chemical Engineers,
1994)]
Dows Fire and Explosion Index FEI (American Institute of Chemical Engineers,
1987; Khan & Abbasi, 1998a)

Checklist (Khan & Abbasi, 1998b)

Accident Sequences Precursor ASP (Holmberg, 1996)

Annex 6 of SEVESO II Directive (La directive Seveso II: Annexe 6, 1997)]
AVRIM2 (Ham et al., 1998)

Accident Hazard Analysis AHI (Khan & Abbasi, 1997b; Khan & Abbasi, 1998a)

Methodologies

Table 5
Application fields of risks analysis methodologies

J. Tixier et al. / Journal of Loss Prevention in the Process Industries 15 (2002) 291303
299

Human

Transport

Table 5 (continued)

Hazard and Operability HAZOP (Kennedy & Kirwan, 1998; Khan & Abbasi, 1998b;
Nicolet-Monnier, 1996; Rogers, 2000; Tweeddale et al., 1992)
Event Tree Analysis ETA (Gadd et al., 1998; Nicolet-Monnier, 1996; Rogers, 2000; IPO Risico Berekening Methodiek IPORBM (Tiemessen & van Zweeden, 1998)
Tiemessen & van Zweeden, 1998]
Failure Mode Effect Analysis FMEA (Khan & Abbasi, 1998b; Nicolet-Monnier,
Quantitative Risk Assessment QRA (Alonso & Gavalda, 1998; Khan & Abbasi,
1996; Rogers, 2000)
1998b; Leeming & Saccomanno, 1994; Oien et al., 1998; Puertas et al., 1998;
Papazoglou et al., 1992; Rogers, 2000)
Fault Tree Analysis FTA (Khan & Abbasi, 1998b;Nicolet-Monnier, 1996; Rogers,
What if? Analysis (Khan & Abbasi, 1998b; Nicolet-Monnier, 1996; Rogers, 2000)
2000)
Action Errors Analysis AEA (Rogers, 2000)
Safety Culture Hazard and Operability SCHAZOP (Kennedy & Kirwan, 1998)
Human Hazard and Operability HumanHAZOP (Kennedy & Kirwan, 1998)
Task Analysis TA (Rogers, 2000)
Manager (Pitblado et al., 1990)
Work Process Analysis Model WPAM (Davoudian et al., 1994)
Process Risk Management Audit PRIMA (Hurst et al., 1996)

Checklist (Khan & Abbasi, 1998b)

Methodologies

300
J. Tixier et al. / Journal of Loss Prevention in the Process Industries 15 (2002) 291303

J. Tixier et al. / Journal of Loss Prevention in the Process Industries 15 (2002) 291303

301

Table 6
Risks analysis methodologies with hierarchisation step
Methodologies with hierarchisation phase
Accident Hazard Analysis AHI (Khan & Abbasi, 1997b; Khan & Abbasi, 1998a)
Annex 6 of SEVESO II Directive (La directive Seveso II: Annexe 6, 1997)
Chemical Runaway Reaction Hazard Index RRHI (Kao & Duh, 1998)
Dows Chemical Exposure Index CEI (American Institute of Chemical Engineers, 1994)
Dows Fire and Explosion Index FEI (American Institute of Chemical Engineers, 1987; Khan & Abbasi, 1998a)
Earthquake safety of structures and installations in chemical industries (Jezler, 1998)
Facility Risk Review (Schlechter, 1996)
Failure Mode Effect Criticality Analysis FMECA (Rogers, 2000
Fire and Explsion Damage Index FEDI (Khan & Abbasi, 1998a)
Hazard Identification and Ranking HIRA (Khan & Abbasi, 1997b; Khan & Abbasi, 1998b)
Instantaneous fractionnal loss index IFAL (Khan & Abbasi, 1998a; Khan & Abbasi, 1998b)
Methodology of domino effects analysis (Dolladille, 1999)]
Methods of potential risk determination and evaluation (Ja ger & Ku hnreich, 1998)
Mond Fire Explosion and Toxicity Index FETI (Khan & Abbasi, 1998a; Khan & Abbasi, 1998b)
Potential domino effects identification (Delvosalle et al., 1998)
Probabilistic Safety Analysis PSA (Khan & Abbasi, 1998b; Papazoglou et al., 1992)
Risk Level Indicators RLI (Oien et al., 1998)
SAATY methodology (Troutt & Elsaid, 1996)
Seqhaz Hazard Mapping SHM Korjusiommi et al., 1998
Toxic Damage Index TDI (Khan & Abbasi, 1998a)

toxicity with a correlation dose/effect and the study

of their incompatibility;
quantities of substances used and stored;
storage characteristics (pressure, temperature,).
Severity of the consequences:
human damage types ( over pressure blast, thermal
flux, toxicity);
equipment damage types;
environmental damage types;
financial loss on equipment or production.
Layout and environment:
distances between dangerous units of an industrial
site;
population density inside and outside the industrial
site.
The probabilistic scales are quantitative and they are
made of variables which are:
occurrence frequency of hazardous events;
incident or accident frequencies, historical knowledge;
consequences probabilities (fatalities, structure damages, ground or water pollution, ).
The various scales are developed for several types of
application according to the particularities of certain
industrial site in order to rank the risk in consistent manner. To generalise hierarchisation scale, it is necessary
to elaborate an exhaustive list of influencing risk factors

from many industrial sites and to take into account

human, environmental and equipment damages.
In fact, the hierachisation phase gives more advanced
data processing in order to rank risks on the studied area
(risk level index). This ranking provides help for
decision-makers
(industrialists
and
competent
authorities).

4. Conclusion
The use of risk analysis methodologies contributes to
the prevention of accidents and to the preparation for
emergency response. This work based on the review of
62 methodologies underlines the difficulty in taking into
account all risks for an industrial site. This paper highlights the different types of input data, methods, output
data and their links. A risk analysis methodology can be
simple and only focus on the identification of hazards
or a combined risk analysis methodology. A combined
risk analysis methodology can be composed of several
simple risk analysis methodologies, with an identification, estimation and hierarchisation phases in order to
obtain a risk level index, for example. The recent development of risk analysis methodologies strives towards
easily applicable methods with a hierarchisation phase,
which is based on specific scales depending on the studied installation. The application fields of methodologies
are industrial site, transport of hazardous goods and
human factors. The human factor risk analysis is often
disconnected with classical risk analysis that is due to
the complexity of human risk analysis. The types of

302

J. Tixier et al. / Journal of Loss Prevention in the Process Industries 15 (2002) 291303

results are recommendations, lists, risk level index, event

frequency and damage probability.
The 62 methodologies identified show that there is not
a uniqueness of methods to realise a risk analysis. On
the contrary, it is necessary to combine several methodologies. The application of these methodologies requires
experience to obtain good results. In fact, the acquired
knowledge through the analysis of the 62 methodologies
can constitute a starting point to elaborate a new methodology. To elaborate a methodology it seems appropriate
to propose an initial draft of methods. Then its application on real industrial sites would improve it. Therefore, its transferability to other cases would be easier.
The whole remarks, presented in this study, aim to
propose a draft of an ideal risk analysis methodology.
First, the studied area must be put in four parts to lead
the risk analysis:

the source term (industrial establishment);

the flux (vector of propagation of accidents);
targets (human, environmental and equipments);
control and management.

These four identified parts must be described in an

exhaustive way and together with their interactions.
Then, the mainspring of risk analysis should permit a
deterministic and probabilistic approach with a hierarchisation phase and finally, the output data could be
of two different types:
qualitative in order to provide recommendations;
quantitative in order to evaluate the main consequences
This methodology presents an overall process of risk
analysis in order to provide some ways of improvement
and help in decision-making.

References
Alonso, C., & Gavalda, J. (1998). A method to determine environmental risk in chemical process industries. In Proceeding from ninth
international symposium loss prevention and safety promotion in
the process Industries (pp. 12191227).
DOWs Fire & Explosion Index Hazard Classification Guide. (1987).
American Institute of Chemical Engineers.
DOWs Chemical Exposure Index Guide. (1994). American Institute
of Chemical Engineers (1st ed.).
Chung, P. W. H., Broomfield, E., & Yang, S. H. (1998). Safety related
questions for computer-controlled plants: Derivation, organisation
and application. Journal of Loss Prevention in the Process Industries, 11, 397406.
Davoudian, K., Wu, J. S., & Apostolakis, G. (1994). The work process
analysis model (WPAM). Reliability Engineering and System
Safety, 45, 107125.
Delvosalle, C., Fievez, C., & Benjelloun, F. (1998). Development of
a methodology for the identification of potential domino effects in
Seveso industries. In Proceeding from ninth international sym-

posium loss prevention and safety promotion in the process industries (pp. 12521261).
Dolladille, O. (1999). Proposition dune me thode danalyse deffet
domino. Pre ventique-se curite , 44, 6270.
Gadd, S. A., Leeming, D. G., & Riley, T. N. K. (1998). Transport
Riskat: The HSE quantified risk assessment tool for toxic and
flammable dangerous goods transport by road and rail in Great Britain. In Proceedings from ninth International symposium loss prevention and safety promotion in the process industries (pp. 308
317).
Ham, K. J. M., van Kessel, H. J. C. M., & Wiersma, T. (1998). Experiences with a safety report according to Seveso II: A pilot project
in the Netherlands. In Proceeding from ninth international symposium loss prevention and safety promotion in the process Industries (pp. 13261340).
Holmberg, J. (1996). Risk follow up by probabilistic safety assessment-experience from a finish pilot study. Reliability Engineering
and System Safety, 53, 315.
Hurst, N. W., Young, S., Donald, I., Gibson, H., & Muyselaar, A.
(1996). Measures of safety management performance and attitudes
to safety at major hazard sites. Journal of Loss Prevention in the
Process Industries, 9(2), 161172.
Ja ger, P., & Ku hnreich, K. (1998). Approach to a systematic determination and evaluation of risk potential. In Proceeding from ninth
international symposium loss prevention and safety promotion in
the process industries (pp. 393403).
Jezler, W. (1998). Earthquake safety of structures and installations in
chemical industry in the context of risk analysis. In Proceeding
from ninth international symposium loss prevention and safety promotion in the process industries (pp. 414421).
Kao, C. S., & Duh, Y. S. (1998). Chemical runaway reaction hazard
index and risk assessment. In Proceeding from ninth international
symposium loss prevention and safety promotion in the process
industries (pp. 965975).
Kennedy, R., & Kirwan, B. (1998). Development of a hazard and
operability-based method for identifying safety management vulnerabilities in high risk systems. Safety Sciences, 30, 249274.
Khan, F. I., & Abbasi, S. A. (1997a). OptHazopan effective and
optimum approach for Hazop study. Journal of Loss Prevention in
the Process Industries, 10(3), 191204.
Khan, F. I., & Abbasi, S. A. (1997b). Accident hazard index: A multiattribute method for process industry hazard rating. Process safety
and environmental protection, 75(4), 217224.
Khan, F. I., & Abbasi, S. A. (1998). Inherently safer design based
on rapid risk analysis. Journal of Loss Prevention in the Process
Industries, 11, 361372.
Khan, F. I., & Abbasi, S. A. (1998a). Multivariate hazard identification
and ranking system. Process Safety Progress, 17(3), 157170.
Khan, F. I., & Abbasi, S. A. (1998b). Techniques and methodologies
for risk analysis in chemical process industries. Journal of Loss
Prevention in the Process Industries, 11, 261277.
Korjusiommi, E., Salo, R., & Taylor, R. (1998). Hazard analysis for
batch processes and for special operations. In Proceeding from
ninth international symposium loss prevention and safety promotion in the process industries (pp. 422431).
Kusiak, A., & Zakarian, A. (1996). Risk assessment of process models.
Computers & Industrial Engineering, 30(4), 599610.
La directive Seveso II: Annexe 6. (1997). edition legislative, bul 230,
69336947.
Lagadec, P. (1980). Politique, risque et processus de de veloppement:
Le risque technologique majeur, The`se de doctorat de tat en
science politique, Universite des Sciences Sociales de Grenoble II.
Larson, N., & Kusiak, A. (1996). Managing design processes: A risk
assessment approach. IEEE Transactions on Systems, Man, and
CyberneticsPart A: Systems and Humans, 26(6), 749759.
Leeming, D. G., & Saccomanno, F. F. (1994). Use of quantified risk

J. Tixier et al. / Journal of Loss Prevention in the Process Industries 15 (2002) 291303

assessment in evaluating the risks of transporting chlorine by road

and rail. Transportation Research Record, 1430, 2735.
Nicolet-Monnier, M. (1996). Integrated regional risk assessment: The
situation in Switzerland. International Journal of Environment and
Pollution, 6(4-6), 441461.
Oien, K., Sklet, S., & Nielsen, L. (1998). Development of risk level
indicators for petroleum production platform. In Proceeding from
ninth international symposium loss prevention and safety promotion in the process industries (pp. 382393).
Papazoglou, I. A., Noivolianitou, Z., Aneziris, O., & Christou, M.
(1992). Probabilistic safety analysis in chemical installation. Journal of Loss Prevention in the Process Industries, 5(3), 181191.
Perhillon, P. (2000). Ele ments me thodiques. Phoebus, la revue de la
su rete de fonctionnement, Lanalyse de risques, 12, 3149.
Pitblado, R. M., Williams, J. C., & Slater, D. H. (1990). Quantitative
assessment of process safety programs. Plant/Operations Progress,
9(3), 169175.
Puertas, I., Sanz, J. C., Vaquero, C., Marono, M., & Sola, R. (1998).
Procedure for the review of quantitative risk assessment of the process industries. In Proceeding from ninth international symposium
loss prevention and safety promotion in the process industries (pp.
283288).
Rasmussen, B., & Whetton, C. (1997). Hazard identification on plant
functional modelling. Reliability Engineering and System Safety,
55, 7784.
Rogers, R.L. (2000). The RASE Project risk assessment of unit oper-

303

ations and equipment, http://www.safetynet.de/EC-Projects/. pp.

1-50.
Sankey, P. D. (1998). An insurers involvement in the risk reduction
process. In Proceeding from ninth international symposium loss
prevention and safety promotion in the process industries (pp.
441450).
Schlechter, W. P. G. (1996). Facility risk review as a mean to addressing existing risks during the life cycle of a process unit, operation
or facility. International Journal of Pressure Vessels and Piping,
66(13), 387402.
Tiemessen, G., & van Zweeden, J. P. (1998). Risk assessment of the
transport of hazardous materials. In Proceeding from ninth international symposium loss prevention and safety promotion in the
process industries (pp. 299307).
Toola, A. (1992). Plant level safety analysis. Journal of Loss Prevention in the Process Industries, 5(2), 119124.
Troutt, M. D., & Elsaid, H. H. (1996). The potential value of SAATYs
eigenvector scaling method for short-term forecasting of currency
exchanges rates. Siam Review, 38, 650654.
Tweeddale, H. M., Cameron, R. F., & Sylvester, S. S. (1992). Some
experiences in hazard identification and risk shortlisting. Journal
of Loss Prevention in the Process Industries, 5(5), 279288.
Yang, S., & Chung, P. W. H. (1998). Hazard analysis and support tool
for computer controlled processes. Journal of Loss Prevention in
the Process Industries, 11, 333345.