Hazard and Risk Assessment Practices Review of Indian

Experience in the Light of International Practices

(Hazard and Risk Assessment Tools have found increasing application within
industries/establishments storing and handling hazardous chemicals in India over the
past decade or so, largely as a result of statutory pressures. A large number of studies
have been carried out both for the purpose of environment clearance and at the behest
of the MoE&F to generate information for use in Off-site Emergency Planning. In the
absence of consensus methodologies and/or statutory guidelines, the experience with
these studies remains mixed at best. This paper reviews the experience gained, and
then goes on to highlight current practices elsewhere with a view to evolve a road-map
and recommendations for improvement. It draws upon the findings of the Review of
Indian Information Base and Review of International Base on Hazard Assessment done
as a part of a World Bank sponsored project for the MoE&F)

1.

Introduction

In India, the impetus for the application of techniques of Hazard and Risk Assessment,
particularly of Quantitative Risk Analysis, in process industry came largely as a result of
statutory pressure. The Ministry of Environment and Forests (MoE&F) notified, in 1989,
the Manufacture, Storage, and Import of Hazardous Chemicals Rules, under the
Environment (Protection) Act. Later, in 1994, the MoE&F notified regulations for
according Environmental Clearance.
These regulations on Environmental Clearance clearly require that Risk Analysis be
carried out for specific categories of projects in support of applications for their
environmental clearance.
This Manufacture Storage & Import of Hazardous Chemicals (MSIHC) Rules call for the
submission of a Safety Report by industries storing or handling certain hazardous
chemicals in excess of defined threshold values. Preparation of the Safety Report
requires first of all that a Preliminary Hazard Analysis be carried out to identify the types
of accidents that could occur, the system elements or events that could lead to a major
accident, the hazards and the safety relevant components. HAZOP is expected to be
carried out for the safety relevant systems or units identified.
Preparation of the Safety Report all requires major inputs from a Risk Analysis study.
Specifically, details of effects-consequence analysis and accident frequency analysis

2
need to be furnished. Therefore, Risk Analysis study is a pre-requisite to preparation of
the Safety Report.
Risk Analysis is a complex exercise and can be carried out to various depths. Therefore,
clear definition of study goals and study requirements is important, as these determine
the depth of a given study. (1)
A Risk Analysis study may have several objectives. Some of the most common ones
are:
To demonstrate that the risks (to public) associated with a process are tolerable as
judged against acceptance criteria. This is a common objective and one that needs
to be often met for reasons of obtaining regulatory or statutory clearance for new
projects. It may be done for either new plant/process or modification to existing
plant/process.
To identify opportunities for risk reduction and evolve suitable measures
To assist in land-use planning
To develop hazard scenarios for the purpose of Emergency Management Planning
Cost Benefit analysis to evaluate various project options as well as measures for risk
reduction

2.

When transport of hazardous chemicals is considered, Transport Risk Analysis may

be used to
Understand influence of material state
Select container and/or additional protective features of container
Select mode of transportation
Choose among alternative routes on relative basis

Review of Indian Experience

There is about a decade of experience in India of carrying out Risk Analysis studies.
Numerous studies have been carried out by a host of companies, large and small, in
both public and private sectors. However, in the absence of any standard prescribed
approach considerable variation may be observed in the approach and results. Quite
naturally, this may lead to concern about the reliability of risk analysis as a decision
making tool. It is important to review the experience gained and draw lessons for future
development so as to put the application of risk analysis on a sounder footing.
For Europe, the Joint Research Centre of the European Community conducted a
benchmarking exercise on risk analysis in 1988-90. The objectives of the study were to
evaluate the state of the art and to obtain estimates of the degree of uncertainty in risk
studies. Eleven teams representing twenty-five organisations participated in the
exercise. The project resulted in a comprehensive overview of currently available
methodologies for chemical risk assessment in Europe and triggered an important
common learning process toward sound analysis procedures. As further fallout, the

3
lessons learnt have influenced the development of national standards and codes for risk
analysis within Europe.
No such bench marking exercise has been conducted for India, but we may draw upon
the European exercise. Moreover, quite recently, a review of the Indian Information Base
in Hazard Analysis was done for the Ministry of Environment and Forests as one portion
of Part 5 of the World Bank funded Industrial Safety and Disaster Prevention Project
(ISDPP-5), [2].

2.1

Review of the Indian Information Base in Hazard Analysis from ISDPP-5

This review was carried out using the results of Hazard Analysis performed by a number
of consultants for the Ministry of Environment and Forests as part of Ministry's
Districtwise Hazard Analysis project. As of 1999, reports for about 25 industrial pockets
throughout India were available, prepared by various consultants on assignment by the
MoE&F. One report from each of the (totally thirteen) consultants that were involved in
these pocketwise studies was reviewed. It was assumed that each consultant would
have used the same approach, methodology and models in all their studies. The MoE&F
had prescribed a micro-methodology for the study to minimise variations in results
arising out of consultants' approach and judgment. Nevertheless, the review indicated a
wide variation in approach and understanding of the subject.
Deficiencies of understanding, among some consultants, were noted in identification of
hazards and failure modes, formulating source terms, dispersion modelling, and in
modelling of vapour cloud explosions. In short, the understanding of effects and
consequence modelling was found to be inadequate.
For example, the findings of the review with respect to modelling of the BLEVE and VCE
scenarios may be considered. The review revealed that many consultants had
considered the sum total of storages for modelling the BLEVE/VCE due to domino or
knock on effects. According to the Manual on Emergency Preparedness for Chemical
Hazards by MoE&F the maximum single storage for each hazardous chemical is
required to be identified and considered for analysis. Even if one failure resulted in
subsequent failures of other vessels the consequences would be determined by the
failure of the largest inventory vessel. The approach adopted by some consultants in
adding up all inventories to determine consequences was therefore incorrect and
resulted in a highly pessimistic outcome.
With regard to the VCE scenario some consultants had not used dispersion models to
calculate the explosive mass in the confined vapour cloud. Instead they had modelled

4
the total released quantity as the explosive mass, which will give unduly pessimistic
results.
Clearly, the observed approaches give rise to great doubts about the knowledge and
understanding in effects and consequences modelling, present with several of the
consultants.

2.2

Review of Other Risk Analysis Reports

Review has also been made from time to time of Risk Analysis reports of a number of
consultants, though no reports have been published owing to the confidential nature of
the information. However, these indicate much the same inadequacies as noted in the
previous sub-section in regard to the identification of hazards and failure modes and the
understanding of effects and consequence modelling. In particular the understanding of
likelihood estimation is found to be particularly lacking.
In fact confusion may be noted right from the stage of identification of accident scenarios
for analysis. While there are several techniques for hazard identification available, the
review of release sources is the most appropriate for risk analysis. Other techniques
such as HAZOP have their own applications but their use towards generating scenarios
for risk analysis is limited. HAZOP and other techniques such as FMEA (Failure Mode
and Effects Analysis) may be used for developing fault trees for determining the failure
likelihood of complex systems and complex failure modes, though their application may
not be essential.
The choice of objectives of the risk analysis study will influence the scenarios to be
considered after hazard identification, the extent of their development and the risk
measures to be computed. Naturally, what is appropriate for a given objective or set of
objectives may not be so when this/these are varied.
Likelihood estimation requires data on equipment failure frequencies and reliability.
Failure frequencies may be classified as generic and those synthesised for a particular
situation, especially more complex systems, by drawing Fault Trees that make use of
generic failure frequencies of components of a system or its subsystems. Generic failure
frequencies are preferred wherever these are available, as these reduce variances
arising out of analyst judgement in the failure frequency estimation. Many consultants do
not seem to understand this distinction.
Instances have also been observed where no fault tree analysis has been done but a
standard fault tree is drawn without indicating the specific frequencies that are to be
used for the given situation. This merely misleads the user of the risk analysis report. In

5
other cases, so-called pathway reliability analysis is talked about, again without
reference to specifics, which serves no useful purpose.
Some others contend that likelihood estimation should not be done at all as there are no
India specific data available and the applicability of data pertaining to installations
abroad to Indian situation is questionable. While there can be no argument about the
need to collect and analyse equipment reliability data specific to India, the argument
about the non-applicability of data of other countries is fallacious. Plant and equipment
used in India and abroad are designed to the same codes and adopting the same Best
Practices. Where, there could be a difference is in the standards of Safety Management,
including operating and maintenance procedures.
It could be argued that the equipment reliability data needs to be modified for variations
that could arise on account of these differences. The impact of these differences could
be of an order of magnitude, and the assessment of the impact of Safety Management
on equipment reliability is an area that merits research. Pending such research, any
modification of frequencies may be contestable. It is our opinion that such discounting is
not necessary, unless it is required specifically to illustrate the influence of improved
management practices on the risk. In any case, the results of risk analysis are not to be
used in an absolute sense but find application only when compared against appropriate
criteria established within the context of the risk analysis methodology to be used. With a
well-defined methodology for calculating risk levels, consistent results can be expected.
The inadequacy of understanding of the risk analysis process, and also possibly
confusion about the objectives of the study, results also in the use of risk measures and
modes of risk presentation which are irrelevant to the objectives of the study and that do
not contribute an improved appreciation of the risks involved. For instance, many
consultants are in the habit of presenting the consequence distances of scenarios on the
plot plan and calling these plots as risk contours. This is incorrect as risk involves also
the consideration of likelihood and not merely consequence alone. In any case, as such
consequence distance plots are circles, and as the consequence distances are available
in the report, no additional information is presented, while the bulk of the report is
increased considerably.
It may be concluded that there is much scope for improvement in the knowledge base
and expertise level concerning risk analysis. In this connection, the development of
guidelines for risk analysis by the regulatory authorities is indicated, as the majority of
such reports are prepared for the purpose of obtaining regulatory approval. Such
guidelines may be drawn up following review of guidelines applied in other countries and

6
discussions with experts on the subject as well as consultants, while following a process
of due review and comments by user industry and the public at large. Such an effort
would go a long way in ensuring that studies carried out are to a minimum acceptable
standard.

3.

International Trends

The Netherlands has been a pioneer in the application of risk analysis as an aid to
decision making and the authorities there have sponsored the development of risk
analysis methodology and tools for risk analysis, which have subsequently been adopted
as standards not only within The Netherlands but also in the European Community and
elsewhere in the world. TNO of The Netherlands has played a major role in these efforts.
The publication of the internationally renowned Yellow, Green, and Red Books is
testimony, [3, 4, 5.].
The Yellow Book is a standard reference for effects modelling. The Green Book is a
standard reference for consequence analysis. The Red Book gives guidance on
Likelihood Estimation, including Event Trees and Fault Trees.
Recently, the Dutch authorities have published the so-called Purple Book, which gives
Guidelines for Quantitative Risk Assessment, [6]. This book pertains to Risk Analysis to
be performed for determining the risks to public (off-site) associated with an industrial
activity involving hazardous chemicals. It lays down the methodology to be adopted for
selection of release sources and activities to be considered for QRA, followed by
specification of the Loss of Containment events to be considered from each such
selected inventory/activity and guidelines for effects and consequence modelling and
risk summation and presentation.
The Purple Book also gives guidance on failure frequencies. These are based largely on
the so-called COVO study of 1981, [7]. However, these appear to be rather low when
compared with figures from other sources that have become available. Newer studies
have been published that report different figures mostly higher for a number of failure
frequencies. It is recognised that more detailed study on the failure frequencies require
to be carried out, especially on the original data sources.
The use of the standard methods prescribed in the above references requires the use of
appropriate software, as manual computations are too involved and cumbersome to be
feasible in any practical situation. TNO and many other agencies have developed
software incorporating the models referenced above. The relevant TNO software are as
follows:

7
EFFECTS4 for effects modelling
DAMAGE for damage analysis
RISKCURVES for risk summation and plotting of risk contours

4.

Risk Criteria

4.1

Need For Risk Assessment Criteria

A risk analysis provides a measure of the risks resulting from a particular facility or
activity. It thus finds application as a decision making tool in situations where judgement
has to be made about the tolerability of the risk posed by an existing/proposed activity.
However, risk analysis produces only numbers, which themselves provide no inherent
use. It is the assessment of those numbers that allows conclusions to be drawn and
recommendations to be developed. The assessment phase of a study is therefore of
prime importance in providing value from a QRA study. The normal approach adopted is
to relate the risk measures obtained to risk acceptance criteria.
As stated above, the value of a QRA study depends on the quality of the assessment
phase of that study and, as the quality of a risk assessment depends on the availability
of appropriate and well-defined risk acceptability criteria, it follows that the value of a
QRA study depends on the quality of the criteria used.
In general one can say that risk from an activity should be judged against the benefits
from the activity, since no activity can claim to be totally risk-free. The reason for
permitting new developments is that the benefits, in economic or social terms, justifies
the risk induced by the activity.
Industrial activities involve risk. In particular, an activity may involve risk of fatality to
people. These may be employees engaged in the activity or members of the public
outside. Where an industrial activity imposes risk of fatality to people outside by the very
nature of the activity undertaken, it becomes the concern of regulators to assess the risk
and evaluate it in terms of risk criteria.
Risk criteria, if they are to be workable, recognise the following:
There is a level of risk that is so high that it is considered unacceptable or intolerable
regardless of the benefits derived from an activity.
There is also a level of risk that is low enough as to be considered negligible.
Levels of risk in between are to be considered tolerable subject to their being
reduced As Low As is Reasonably Practicable (ALARP). (The meaning of ALARP is
explained in the following sub-section.)

8
The above is the formulation of the, now well-established, three tier structure of risk
criteria and risk control. (See Fig 2).
The risk criteria simply attempt to establish whether is risk is tolerable. Below is a list of
words generally in use and their meaning.
ACCEPTABLE RISKS: Since risks in general are unwelcome no risk should be called
acceptable. It might be better to say that the activity may be acceptable generally, but
the risks can only ever be tolerable.
TOLERABLE RISKS are risks the exposed people are expected to bear without undue
concern. A subtle difference is made out here between Acceptable Risks and Tolerable
Risks though these terms are sometimes used interchangeably.
NEGLIGIBLE RISKS are risks so small that there is no cause for concern and there is no
reason to reduce them.

4.2

The ALARP Principle

The ALARP (As Low As is Reasonably Practicable) principle seeks to answer the
question What is an acceptable risk? The definition may be found in the basis for
judgement used in British law that one should be as safe as is reasonably practicable.
Reasonably practicable is defined as implying that a computation must be made in
which the quantum of risk is placed on scale and the sacrifice involved in the measures
necessary for averting the risk (whether in money, time, or trouble) is placed on the
other, and that, if it be shown that there is a gross disproportion between them risk
being insignificant in relation to the sacrifice the defendants discharge the onus upon
them.
In other words risks are only tolerable provided that it can be demonstrated that all
reasonably practicable measures have been implemented to reduce risks. A reasonably
practicable risk reduction measure is one where the costs of implementation are not
grossly disproportionate to the risk reduction benefits achieved.
Separate sets of criteria need to be applied for different purposes. The criteria that a
regulatory authority would apply for grant of environmental clearance to a project are
different from those that need be applied for evaluation of employee risks. Again, criteria
suitable for assessing transportation risks are different.

4.3

Risk Criteria in Use

Several countries that have adopted the use of risk analysis as an aid to decision
making in evaluating the environmental impact of industrial projects have evolved criteria
suitable to their requirements. For India, no such criteria have been developed. There is
a need for specifying risk assessment criteria in conjunction with the development of
standard methodologies for risk analysis so as to permit objective decision-making using
the results of a risk analysis study.
The importance of a standard methodology to be applied in conjunction with the risk
criteria cannot be overemphasized. Risk criteria are analogous to specifications applied
to judge the quality of a product, for example specification of flash point or density for
petroleum products. Like with other specifications, test methods to generate the
numbers that will be compared with the criteria are required. Quantitative risk analysis
involves many judgements and assumptions that will impact the resulting risk estimates.
Hence, risk criteria should be established within the context of the risk analysis
methodology to be used.
In the absence of official criteria for India, appropriate Risk Criteria for such studies were
drawn up after a review of official criteria in use in other countries. These risk criteria and
the criteria framed for studies in the Indian context are discussed below.

Individual Risk Criteria

General
Individual Risk is defined as the risk of death and is expressed the probability of fatal
injury per year to a hypothetical individual assumed to be continuously present at a
specific location. This kind of result is suitable for presentation in the form of iso-risk
contours.
Once a risk has been defined which members of the public are expected to tolerate from
hazardous activities over which they have no control, it should not matter whether this
risk comes from a single plant, an industrial complex or a transport accident.
Workers involved in the hazardous activities benefit from it directly and may be expected
to tolerate a higher level of risk than workers on neighbouring plants and the public in
general. An order of magnitude higher risk is proposed as tolerable for workers on the
hazardous plant. Given below is a summary of official criteria developed in various
countries.
Table 1

Official Individual Risk Criteria Of Select Countries

10
AUTHORITY

INTOLERABLE RISK

NEGLIGIBLE RISK

(per year)

(per year)

VROM, The Netherlands (New Plants)

10-6

10-8

VROM, The Netherlands (Existing plants or

10-5

10-8

10-5

10-6

10-4

10-6

10-5

10-6

10-5

Not used

combined new plants

Environmental Protection Authority, Western
Australia (New Plants)
Health and Safety Executive, UK (Nuclear
power station)
Health and Safety Executive, UK (New housing
near existing plants)
Hong Kong Government (New plants)

The Dutch Government criteria are recognised to be quite strict and have caused
considerable difficulty to their industry in meeting them.

Individual Risk Criteria Applied in the Indian Context

Individual Risk of death to members of the public outside the plant boundaries:
Greater than 10-5 per year is intolerable risk
Lower than 10-6 per year is negligible risk.
An example of the application of these criteria is presented in Fig 3.

Societal Risk Criteria

Society usually judges accidents that result in multiple fatalities more harshly than
multiple accidents that cause fewer fatalities per accident.
Societal risk criteria are after expressed as lines on an F-N curve, showing the
cumulative frequency (F) of accidents involving N or more fatalities. This results in
control over not only the average number of fatalities from all sizes of accident, but also
the risk of catastrophic accidents killing many people at once. Societal risk evaluation is
particularly important for certain situations whose special features are not recognised by
Individual Risk criteria e.g.

Transport activities, which spread their risks over a constantly changing population
along the routes

Large industrial complexes that expose many people over a wide area

Toxic releases that may effect very large number of people

Societal Risk criteria have not been as widely used as Individual Risk criteria because
the concepts and calculations involved are much more difficult. However, their value is

11
becoming recognised, especially for transport activities, but also as complementary to
Individual Risk criteria in general. Table 2 below is a summary of official criteria
developed in various countries. The Dutch Governments criteria are evidently quite
strict. These criteria are illustrated with the aid of Fig 4.
The slope of the F-N curve is a measure of "risk aversion" or the aversion with which
society views accidents with larger number s of fatalities. A slope less than -1 indicates
that society is increasingly averse to multiple fatality incidents as the number of fatalities
increases. A slope equal to -1 indicates that society views multiple fatality accidents
neutrally with respect to only the numbers of fatalities. (However, this is not so with the
frequency of such incidents. Even for slope = -1, the tolerable frequency of accidents
causing different numbers of fatalities is in the inverse ratio of fatalities caused, which
indicates the increased repugnance with which larger numbers of fatalities are viewed.)
The Dutch Government's criteria with slope of 2 indicate extreme risk aversion.
Table 2

Official Societal Risk Criteria of Select Countries

AUTHORITY

VROM,

The

FN CURVE

INTOLERABLE

NEGLIGIBLE

LIMIT

SLOPE

INTERCEPT WITH

INTERCEPT

ON N

N=1

WITH N=1

-3

10-5

Netherlands

-2

10

Government

-1

10-3

-1

10-1

10-4

(New plants)
Hong

Kong

1000

(New plants)
HSE UK (Existing Plants)

Societal Risk Criteria Applied in the Indian Context

Table 3 presents the societal risk criteria judged as suitable for use in Indian studies. It
drew on the criteria applied by the Hong Kong Government for intolerable risk as suited
to our conditions. It may be noted that starting point of the intolerable risk criterion line is
the same as that used in The Netherlands, only the slope of the Dutch criterion is 2,
indicating again their extreme risk aversion. The criterion for negligible risks was kept
low: though not as low for the Dutch case on account of their extreme risk aversion. The
starting point of the negligible risk F-N criterion line is one order of magnitude lower than
the Dutch line and it is defined two orders of magnitude lower than for UK. This implies
that the ALARP zone where industries have to take all practicable measures for risk
reduction is correspondingly larger. Fig 5 presents an example of the application of
these criteria.

12
Table 3

Recommendations On Societal Risk Criteria

FN

CURVE INTOLERABLE

SLOPE

Existing and new plants

5.

-1

NEGLIGIBLE

LIMIT

INTERCEPT

INTERCEPT

ON N

WITH N=1

WITH N=1

10-3

10-6

Conclusion

It may be concluded that there is much scope for improvement in the knowledge base
and expertise level concerning risk analysis. In this connection, the development of
guidelines for risk analysis by the regulatory authorities is indicated, as the majority of
such reports are prepared for the purpose of obtaining regulatory approval. Such
guidelines may be drawn up following review of guidelines applied in other countries and
discussions with experts on the subject as well as consultants, while following a process
of due review and comments by user industry and the public at large. Such an effort
would go a long way in ensuring that studies carried out are to a minimum acceptable
standard. Research on failure frequency data, particularly in Indian conditions, is
necessary to enhance confidence in the results when these are to be used in a
judgemental way in conjunction with risk assessment criteria to decide on the tolerability
of risk from a proposed or existing activity. There is a need for specifying risk
assessment criteria for India, in conjunction with the development of standard
methodologies for risk analysis so as to permit objective decision-making using the
results of a risk analysis study. Authorities in government, industry, legal experts, NGOs
and risk analysis experts need to be involved in the process of specifying suitable
criteria for India.

References
[1] Centre for Chemical Process Safety. Guidelines for Chemical Process Quantitative
Risk Analysis, American Institute of Chemical Engineers, 1989
[2] TNO Institute of Environmental Sciences, Energy Research and Process Innovation
& KLG-TNO Safety Technology Ltd. Final Report on Feasibility Studies for Industrial
Pocketwise Risk Assessment, Task B.1 (Industrial Safety and Disaster Prevention
Project Part 5) to the Ministry of Environment and Forests, New Delhi, March 1999.
[3] Committee for the Prevention of Disasters. Methods for the calculation of physical
effects (the 'Yellow Book'), Third Edition. The Hague: SDU, 1997
[4] Committee for the Prevention of Disasters. Methods for the calculation of damage
(the 'Green Book'), The Hague: SDU, 1990

13
[5] Committee for the Prevention of Disasters. Methods for determining and processing
probabilities ('Red Book'), Second Edition. The Hague: SDU, 1997
[6] Committee for the Prevention of Disasters. Guidelines for Quantitative Risk
Assessment (the 'Purple Book'), The Hague: SDU, 1999
[7] Cremer and Warner Ltd. Assessment of Industrial Risks in the Rijnmond Area (the
COVO study), August 1979, published by D.Reidel Publishing Company, Dordrecht,
Holland, 1982.

14

Identification of
Accident Scenarios

Frequency of Accident
Scenarios

Calculation of Physical
Effects

Probability of Physical
Effects

Calculation of Damage

Probability of Damage

Calculation of Overall
Frequency

Risk Quantification

Risk Evaluation

Risk Reduction

Fig 1 Risk Analysis Process

Risk cannot be
justified on any
grounds

Intolerable Region

Intolerable level
Tolerable only if risk reduction in
impracticable or if its cost is
grossly disproportionate to the
improvement gained

The ALARP region

(Risk is undertaken
only if a benefit is
desired)

Tolerable if cost of reduction

would exceed the improvement
gained

Broadly acceptable
region
(No need for detailed
working to
demnstarate ALARP)

Negligible risk

Fig 2 Levels of Risk and the ALARP criterion

15

2000.00

1500.00
1 E-5/per year

1000.00

500.00
1 E-5/per year

1 E-6/per year

0.00

-500.00

1 E-6/per year

-1000.00

-1500.00

-2500m

Scale:

-2000m

-1500m

-1000m

-500m

0m

1:27948

Fig. 3 Iso-Risk Contours

500m

1000m

Fig 4 Official Societal Risk Criteria of Selected Co

Expected frequency of >= N prompt fatalities

per year

1.00E+00
1.00E-01
1.00E-02
1.00E-03
HSE UK Intolerable Risk
1.00E-04
Hong Kong Intolerable
Risk Line

1.00E-05
1.00E-06

Dutch Intolerable
Risk Line

1.00E-07
HSE UK Intolerable Risk
1.00E-08

Dutch Negligible
Risk Line

1.00E-09
1

10

100

1000

10000

100000

Number of prompt fatalities, N

d:\word2pdf\84722.doc