Risk ID & Assessment Training Course DNV Consulting

Module 11: Risk ranking

MODULE 11

RISK RANKING

P:\2004 Contracts\21506545 PetroVietnam HAZOP+QRA Course\CD-ROM\Word files\Module11.doc © 2004 DNV. All rights reserved
Risk ID & Assessment Training Course DNV Consulting
Module 11: Risk ranking Page i

CONTENTS

11. RISK RANKING .........................................................................................................1

11.1 RISK RANKING HAZARD IDENTIFICATION STUDIES..................................1

11.2 SEVERITY INDEX...............................................................................................1

11.3 FREQUENCY INDEX ..........................................................................................2

11.4 RISK INDEX.........................................................................................................5

11.5 TIMING OF THE RISK RANKING PROCESS ....................................................6

11.6 THE DELPHI TECHNIQUE .................................................................................6

P:\2004 Contracts\21506545 PetroVietnam HAZOP+QRA Course\CD-ROM\Word files\Module11.doc © 2004 DNV. All rights reserved
Risk ID & Assessment Training Course DNV Consulting
Module 11: Risk ranking Page 1

11. RISK RANKING

11.1 RISK RANKING HAZARD IDENTIFICATION STUDIES

When a HAZOP team discusses an issue whose consequence is related to safety,

environmental consequence, or material damage they must use qualitative risk assessment.
Once they accept that an event may occur, they must answer the questions "How likely is it?"
and "How serious would it be ?" Unless they use calculation, they will apply qualitative risk
assessment to decide what to do.

During the study a consensus will be reached on whether the risk is acceptable or
unacceptable, or the problem will be referred for further analysis. This consensus may be
reached after discussion by the team and a judgment made or alternatively the 'Delphi Method'
can be used to aid decision making. This is discussed later.

Some more sophisticated uses of this method are possible, either during a HAZOP study, or in
follow-up of unresolved issues. These techniques are discussed below.

Simple Risk Index and Delphi Method

If a group of people are asked to discuss a particular hazard scenario and decide whether the
risk is acceptable, it is often difficult for them to draw a conclusion from an unstructured
discussion.

One method to overcome this problem is to use a simple, subjective risk index. This may be
done as described in the following sections. The method described in this section is that most
commonly adopted (for example in BS 8444 Part 3 1996 Risk Management). However, the
technique is purely subjective and other scales are often used.

11.2 SEVERITY INDEX

This index varies between 1 and 4, and ranks the severity (or consequences) of the event
according to the following descriptions:

1. Minor
2. Marginal
3. Critical
4. Catastrophic

Consequences may be better described as ‘potential losses’ since these consider any losses
which are included in the ground rules for the hazard identification study. Examples of losses
are:

• Safety issues (injury or fatality)

• Environmental effects/damage
• Operability losses, e.g. loss of revenue, production loss, etc)
• Loss as the result of project delay/commercial project losses

P:\2004 Contracts\21506545 PetroVietnam HAZOP+QRA Course\CD-ROM\Word files\Module11.doc © 2004 DNV. All rights reserved
Risk ID & Assessment Training Course DNV Consulting
Module 11: Risk ranking Page 2

Once again some further expansion of these categories is necessary as follows:

Index Description Definition Value of Loss

1 Minor Possible single minor injury/minor plant damage < £10,000
2 Severe Serious injury/significant plant damage £10,000 - £100,000
3 Major Severe injury or single fatality/extensive plant damage £100,000 - £1m
4 Catastrophic Multiple fatalities/complete loss of plant >£1m

Where more than one final event outcome is possible, in order to simplify the risk index
process it is easiest to use the index for the worst likely severity event. In the case of the
road tanker accident discussed in Section 11.3 this would probably be the consequences of a
pool fire around the tanker since fireballs in this situation tend to be very rare events. The
actual consequences will depend on the particular circumstances, for example how much time
is available for people to move away from the danger area and therefore prevent injury or
fatality.

11.3 FREQUENCY INDEX

This index is a number typically between 1 and 6 which ranks the likelihood of the event,
according to the following descriptions:

1. Incredible
2. Improbable
3. Remote
4. Occasional
5. Probable
6. Frequent

This description can be further aided by the following definitions:

Index Indicative Frequency Definition

(per year)
1 < 10-6 The event is extremely unlikely to occur It can be assumed that it
will not occur.
2 10-4 - 10-6 The event is unlikely to occur. It can be assumed that it will only
occur exceptionally.
3 10-2 - 10-4 The event is likely to occur some time in the system life cycle. It can
reasonably be expected to occur a few times.
4 10-1 - 10-2 The event is likely to occur several times. It can be expected to occur
several times.
5 1 - 10-1 The event will occur several times. It can be expected to occur
frequently.
6 >1 The event will be continually experienced.

The indicative frequencies shown are not fixed and can be changed to suit particular situations.

It is important to realise that the Frequency Index selected is the likelihood of the final event
not the initial event and is therefore dependent on the Severity index selected. To gain a

P:\2004 Contracts\21506545 PetroVietnam HAZOP+QRA Course\CD-ROM\Word files\Module11.doc © 2004 DNV. All rights reserved
Risk ID & Assessment Training Course DNV Consulting
Module 11: Risk ranking Page 3

qualitative estimate of frequency it may therefore help by imagining or even drawing an event
tree to help the team determine the likelihood of the final event.

For example, Figure 11.1 shows an event tree for a petrol road tanker collision. The initial
event is the collision, but there are a number of events which have to occur depending on the
consequence being considered.

If the final outcome being considered is a fireball then the events which must occur are the
initial event (tanker collision), there must be a leak of petrol, there must be a source of ignition
(and therefore a fire) and the fire must be out of control. In this case the initial event is the
accident but the Frequency Index selected should be for the final event i.e. the fireball.

P:\2004 Contracts\21506545 PetroVietnam HAZOP+QRA Course\CD-ROM\Word files\Module11.doc © 2004 DNV. All rights reserved
Risk ID & Assessment Training Course DNV Consulting
Module 11: Risk ranking Page 4

Figure 11.1 Event Tree for Gasoline Tanker Collision

Leak size Ignition Fire Outcome

controlled

Delayed Fire in drains etc.

Large Early Pool fire round tanker

No None

Yes Pool fire round tanker

Collision Yes

Small No Tanker Fireball

No None

None None

P:\2004 Contracts\21506545 PetroVietnam HAZOP+QRA Course\CD-ROM\Word files\Module11.doc © 2004 DNV. All rights reserved
Risk ID & Assessment Training Course DNV Consulting
Module 11: Risk ranking Page 5

11.4 RISK INDEX

The risk index is then simply defined by multiplying together the Frequency index and Severity
index. With this ranking system this will yield a number between 1 and 24.

Refer to the following table and read off the Priority Rating:

6 24 18 12 6
L
I 5 20 15 10 5
K
E 4 16 12 8 4
L
I 3 12 9 6 3
H
O 2 8 6 4 2
O
D 1 4 3 2 1

4 3 2 1

SEVERITY/CONSEQUENCE

The level of risk is indicated by the following shading:

Intolerable 13-24
Tolerable 5-12
Negligible 1-4

The priority rating determines the target time-scale for initiating corrective action appropriate
to the workplace.

An index value between 1 and 4 will normally yield a negligible risk, although if something can
be simply done to reduce the index further, this should be carried out.

An index between 5 and 12 would be a tolerable level of risk. Normally the project manager
should aim to reducing the index to 4 or below. This may typically be justified on a cost-
benefit basis, either subjectively or by detailed cost benefit calculation.

If the Severity index is a large number, attention should be paid to mitigating the consequence
to make the scenario less severe. An index above 12 would indicate a situation which either
requires drastic change, or preferably a different process. Action must be taken to reduce the
risk to tolerable levels, and where reasonably practicable to reduce them to negligible levels.
If there is no obvious way of reducing the risk at reasonable costs then it may be would be
normal to approach this situation by changing to a quantitative approach, and using the more
powerful techniques available to direct the re-design of the system. A detailed cost benefit

P:\2004 Contracts\21506545 PetroVietnam HAZOP+QRA Course\CD-ROM\Word files\Module11.doc © 2004 DNV. All rights reserved
Risk ID & Assessment Training Course DNV Consulting
Module 11: Risk ranking Page 6

analysis may be necessary to aid the decision making process. However the use of QRA
should not be taken lightly since the QRA study itself may entail significant costs.

11.5 TIMING OF THE RISK RANKING PROCESS

It has been found by experience that risk ranking is best done after the hazard identification
process. Risk ranking alongside the brainstorming tends to cause confusion and disagreement
and detracts from the most important part of the process, i.e. identification of the hazards.

Risk ranking is also a subjective process which is very dependent on the team. It does not
produce numbers which can be used as a risk standard and therefore users should be very wary
of using the technique to compare risks on a company-wide basis. It works best when used in
isolation for one particular project and where the team are having difficulty deciding if the
existing risk is acceptable.

The most powerful use of the technique is to risk rank before and after recommendations are
made. In that way the effect of the recommendations on the risk are clearly demonstrated. If
a team considers a recommendation is not necessary (i.e. the existing safeguards are adequate)
then it can be argued that there is no need to formally determine a risk index. This is an
entirely reasonable approach but its adoption may depend on company policy.

11.6 THE DELPHI TECHNIQUE

When using subjective methods such as the Risk Index described above, opinions of a group
of people can be very different, often simply because of the different background and
experience of the people.

The Delphi Technique uses this difference of subjective judgment to advantage. The approach
is best carried out by 4 or more people, so is ideally suited to the HAZOP team environment.
The steps in the procedure, applied to the Risk Index method are as follows:

1. A short discussion of the problem to ensure everyone is reviewing the same scenario,
and the Severity of the same outcome.

2. The individuals decide and record their own judgments of Frequency and Severity
Indices, and Risk Index.

3. The group leader takes these values from each member, and lists them.

4. If there is close agreement on the numbers, the average value is taken.

5. If there is a significant difference between the judgment of any two members, the one
whose index is furthest from the average is asked to justify it. If he has taken account
of a factor not known to the other members, this should be identified and discussed.
All are then asked to review their judgments and change them if they wish to. If again
there is a major discrepancy from a single member, the average is calculated excluding
the different view.

P:\2004 Contracts\21506545 PetroVietnam HAZOP+QRA Course\CD-ROM\Word files\Module11.doc © 2004 DNV. All rights reserved