Purple Book 2

Guideline for quantitative
risk assessment
'PurpIe book'
CPR 18E
Part one: Establishments
dr. P.A.M. Uijt de Haag
dr. B.J.M. Ale
RVIM
Preface
PREFACE
This report documents the methods to calculate the risks due to dangerous substances in the
Netherlands using the models and data available. Calculation of the risks relates, on the one hand,
to stationary installations and, on the other, to transport and related activities.
The report consists of two parts. Part 1, describing the methods to calculate the risks of
stationary installations, was written by the National Institute of Public Health and the
Environment (RIVM) under a supervisory committee of representatives from the subcommission
on Risk Evaluation of the Committee for the Prevention of Disasters (CPR-RE). Part 2, drawn
up under the responsibility of the Ministry of Transport, Public Works and Water Management,
describes the calculation of the risks connected with the transport of dangerous goods, based on
the approach developed in accordance with the Ministry of Housing, Spatial Planning and the
Environment and set down in the last few years in various commissions.
Although the report describes the present-day calculation methods (in practice, no better
methods are currently available), discussions on a number of subjects in the supervisory
committee led to the conclusion that additional research would be necessary to guarantee the
quality of the calculation methods in the future. Three subjects for study were indicated:
A. The failure frequencies of stationary installations. Failure frequencies are based on the so-
called COVO study from 1981. Additional failure frequencies have been determined in various
studies carried out for the Dutch government over the years. Recently, new studies have been
published, reporting different figures - mostly higher - for a number of failure frequencies. A
more detailed study on the failure frequencies will be carried out, concentrating especially on
the original data sources.
B. The meteorological model. Dispersion calculations are carried out as part of the risk analyses
using generally accepted meteorological models and the corresponding meteorological data.
The national model used in air pollution calculations has recently been adapted to include new
insights. At the moment, meteorological statistics are not sufficiently available to apply this
new meteorological model to risk analyses. The relevance of the new model to risk analyses
should be ascertained; furthermore, the consequences which the new model, including the
model parameters, could have on the results of calculating risks should be examined. The study
on these consequences will be started up in the short term.
C. Differences in risk calculations for transport and for stationary installations. The method to
calculate the risks of transporting dangerous goods is comparable to the calculation method
applied to stationary installations. During the last few years, the basic principles of risk
analysis have been discussed and established with the parties involved. Since developments in
the risk calculation methods for transport and stationary installations were separate, several
differences exist between the basic principles in risk calculations for transport and for
stationary installations. These differences relate, among other aspects, to the frequency of
catastrophic failure of tank wagons relative to stationary tanks and to certain loss of
containment scenarios.
Preface
The Committee for the Prevention of Disasters considers it important to have reliable risk
calculations for stationary installations and for transport of dangerous goods; these should, as far
as possible, also be founded on similar basic principles. It is therefore advisable to analyse the
basic principles of the calculation methods and to study the consequences of removing the
differences in the calculation methods. Both the Ministries mentioned above can then decide
whether these differences should actually be reduced.
The discussions show that the methods of risk analysis are still being further developed. The
Committee for the Prevention of Disasters is pleased that with the publication of this report a
substantial contribution will be made to the further development of this risk analysis instrument.
The Committee thanks the government experts, research institutes and industry for their
contributions. The Committee for the Prevention of Disasters is convinced that the report will
be of great value for all those dealing with risk analysis and risk management.
Table of Contents 0.1
TABLE OF CONTENTS
1. INTRODUCTION
2. SELECTION OF THE INSTALLATIONS FOR THE QRA
2.1 INTRODUCTION.................................................................................................................................................................................2.1
2.2 EXCLUSION OF PARTICULAR SUBSTANCES.............................................................................................................................2.2
2.3 THE SELECTION METHOD..............................................................................................................................................................2.2
2.3.1 Definition of installations in an establishment ......................................................................................2.4
2.3.2 Calculation of the indication number, A.....................................................................................................2.4
2.3.3 Calculation of the selection number, S......................................................................................................2.10
2.3.4 Selection of installations......................................................................................................................................2.11
2.3.5 Specific problems.....................................................................................................................................................2.11
APPENDIX 2.A PROCEDURE TO ASSESS THE OBLIGATION TO MAKE A SAFETY REPORT...........................................2.15
APPENDIX 2.B AN EXAMPLE CALCULATION............................................................................................................................2.22
APPENDIX 2.C COMMENTARY......................................................................................................................................................2.28
3. LOSS OF CONTAINMENT EVENTS
3.1 INTRODUCTION.................................................................................................................................................................................3.1
3.2 LOSS OF CONTAINMENT EVENTS AT ESTABLISHMENTS......................................................................................................3.1
3.2.1 Stationary pressurised tanks and vessels..................................................................................................3.2
3.2.2 Stationary atmospheric tanks and vessels................................................................................................3.4
3.2.3 Pipes......................................................................................................................................................................................3.7
3.2.4 Pumps..................................................................................................................................................................................3.8
3.2.5 Heat exchangers.............................................................................................................................................................3.9
3.2.6 Pressure relief devices...........................................................................................................................................3.10
3.2.7 LOCs for storage in warehouses...................................................................................................................3.11
3.2.8 Storage of explosives..............................................................................................................................................3.11
3.2.9 Transport units in an establishment...........................................................................................................3.12
APPENDIX 3.A COMMENTARY......................................................................................................................................................3.17
4. MODELLING SOURCE TERM AND DISPERSION
4.1 INTRODUCTION.................................................................................................................................................................................4.1
4.2 PROPERTIES OF SUBSTANCES......................................................................................................................................................4.2
4.3 OUTFLOW MODELS..........................................................................................................................................................................4.2
4.4 REPRESSION FACTORS...................................................................................................................................................................4.5
4.4.1 Blocking systems.........................................................................................................................................................4.5
4.4.2 Other repression systems.....................................................................................................................................4.6
4.5 POOL EVAPORATION........................................................................................................................................................................4.6
4.6 VAPOUR CLOUD DISPERSION.......................................................................................................................................................4.7
4.6.1 Coupling outflow and vapour cloud dispersion...................................................................................4.7
4.6.2 Modelling the vapour cloud dispersion...................................................................................................4.10
4.6.3 Release inside a building......................................................................................................................................4.11
4.6.4 Fires and plume rise................................................................................................................................................4.13
4.7 IGNITION4.13
4.7.1 Direct ignition..............................................................................................................................................................4.13
4.7.2 Delayed ignition.........................................................................................................................................................4.15
4.7.3 Substances both toxic and flammable........................................................................................................4.15
4.8 EFFECTS OF IGNITION OF A VAPOUR CLOUD........................................................................................................................4.16
4.9 RUPTURE OF VESSELS..................................................................................................................................................................4.16
4.10 METEOROLOGICAL DATA........................................................................................................................................................4.17
APPENDIX 4.A MODEL TO CALCULATE THE PROBABILITY OF DELAYED IGNITION...................................................4.19
APPENDIX 4.B METEOROLOGICAL DATA.................................................................................................................................4.21
5. MODELLING EXPOSURE AND DAMAGE
5.1 INTRODUCTION.................................................................................................................................................................................5.1
5.2 DAMAGE MODELLING....................................................................................................................................................................5.1
5.2.1 Probit functions............................................................................................................................................................5.1
5.2.2 Toxic exposure...............................................................................................................................................................5.3
5.2.3 Fire..........................................................................................................................................................................................5.6
5.2.4 Pressure effects for a vapour cloud explosion.......................................................................................5.8
5.3 POPULATION.......................................................................................................................................................................................5.8
5.3.1 Survey of the population present ...................................................................................................................5.8
5.3.2 Fraction indoors and outdoors........................................................................................................................5.10
APPENDIX 5.A COMMENTARY......................................................................................................................................................5.11
6. CALCULATION AND PRESENTATION OF RESULTS
6.1 INTRODUCTION.................................................................................................................................................................................6.1
6.2 CALCULATION OF THE INDIVIDUAL RISK AND THE SOCIETAL RISK................................................................................6.1
6.2.1 Definition of the grid.................................................................................................................................................6.1
6.2.2 Individual Risk calculation....................................................................................................................................6.2
6.2.3 Societal Risk calculation.........................................................................................................................................6.4
6.2.4 Definition of ignition events for flammable substances..................................................................6.6
6.2.5 Probability of death, Pd, and fraction of deaths, Fd, for toxic substances....................6.10
6.2.6 Probability of death, Pd, and fraction of deaths, Fd, for flammables................................6.13
6.3 PRESENTATION OF THE RESULTS..............................................................................................................................................6.16
APPENDIX 6.A PROBABILITY THAT THE GRID POINT IS COVERED BY THE CLOUD, P
CI
............................................6.18
APPENDIX 6.B SAMPLE CALCULATION OF THE INDIVIDUAL RISK AT A GRID POINT..................................................6.21
7. QUANTITATIVE ENVIRONMENTAL RISK ANALYSIS
8. THE USE OF NEW MODELS IN A QRA
9. UNCERTAINTY IN A QRA
9.1 INTRODUCTION.................................................................................................................................................................................9.1
9.2 SOURCES OF UNCERTAINTY..........................................................................................................................................................9.1
9.2.1 Starting points................................................................................................................................................................9.1
9.2.2 Models.................................................................................................................................................................................9.2
9.2.3 Parameter values...........................................................................................................................................................9.2
9.2.4 Use of the model..........................................................................................................................................................9.3
9.3 QUANTIFICATION OF UNCERTAINTIES......................................................................................................................................9.3
10. REFERENCES
GLOSSARY
SYMBOLS
Introduction 1.1
1. INTRODUCTION
A Quantitative Risk Assessment (QRA) is a valuable tool for determining the risk of the use,
handling, transport and storage of dangerous substances. QRAs are used to demonstrate the risk
caused by the activity and to provide the competent authorities with relevant information to
enable decisions on the acceptability of risk related to developments on site, or around the
establishment or transport route.
If the results of a QRA in the decision-making process are to be used, they must be verifiable,
reproducible and comparable. These requirements necessitate QRAs made on the basis of similar
starting-points, models and basic data. Ideally, differences in QRA results should only arise from
differences in process- and site-specific information. A number of documents for attaining
comparability in the QRA calculations have been published over the years. The Committee for
the Prevention of Disasters (CPR) has issued three reports describing the methods to be used in a
QRA calculation, namely the Red Book, the Yellow Book and the Green Book. The Red
Book, describing the methods for determining and processing probabilities, is to be used to
derive scenarios leading to a loss of containment event [CPR12E]. The Yellow Book describes
the models to determine the outflow and dispersion of dangerous substances in the environment
[CPR14, CPR14E], and finally, the Green Book describes the impact on humans of exposure to
toxic substances , heat radiation and overpressure [CPR16].
All three books provide the scientific information to be used in a QRA on the basis of present-
day knowledge. However, this information is not sufficient to carry out a complete QRA
calculation. Additional information is needed, for example, information related to policy
decisions and data for which adequate scientific knowledge is not available (yet). Usually,
standard values for this type of data are set by consensus following discussions between
representatives from industry, the competent authorities and the central government. The
outcome of these discussions has been published in a number of documents (e.g.[KO 9, KO 12,
KO 20-2, KO 24-2, IPO]). However, the large collection of documents issued over the years,
with documents sometimes superseding one another, has called up a need to merge them all into
one report, making use of experiences gathered in conducting QRA analyses. The outcome then is
this report, Guideline for Quantitative Risk Assessment, in which all necessary starting-points
and data needed to perform a QRA calculation are recorded.
The report is organized in the same way that a QRA calculation is performed, i.e. starting with
the selection of installations and the definition of loss of containment events, followed by
dispersion and effect calculations, and the presentation of the results.
The selection of installations is described in Chapter 2. Since the total number of installations in
an establishment can be very large and not all installations contribute significantly to the risk, it is
not worthwhile to include all installations in the QRA. Therefore a selection method is given to
indicate the installations that contribute most to the risk.
The loss of containment events are defined in Chapter 3. Generic loss of containment events and
failure frequencies are defined for a number of standard installations like storage tanks, transport
units, pipelines and loading equipment. Normally, generic values should be used in the QRA
calculation; however, it is possible to use site-specific information so as to modify loss of
containment events and failure frequencies.
Introduction 1.2
Although models to calculate the outflow and the dispersion of dangerous substances are
extensively described in the Yellow Book [CPR14E], a number of topics are not covered, like
the influence of repression systems on the outflow and the dispersion of dangerous substances,
time-varying releases and the ignition of flammable clouds. A need for standard values for the
location and direction of the release and for meteorological parameters is also felt. These and other
model aspects are therefore investigated in Chapter 4.
The effects of toxic substances, fires and explosions on humans are described in Chapter 5. The
information in this chapter is largely based on the Green Book [CPR16]. However, Chapter 5
also describes how the protection of people staying indoors should be accounted for in the QRA
calculation. Furthermore, some guidance is given on surveying a population near the activity
involving dangerous substances.
To illustrate the computation of both the individual and societal risks, a calculation method is
outlined in Chapter 6. This chapter is intended to demonstrate the principles of a QRA
calculation; it does not give a complete description of an established set of calculation rules.
Guidelines on the presentation of the individual and societal risks are also given.
Finally, Chapters 7 - 9 cover several aspects related to a QRA study. Chapter 7 focuses on the
environmental risk analysis and outlines the use of the model, PROTEUS. Chapter 8 goes into
the subject of the use of new models and Chapter 9 considers some aspects of uncertainty in a
QRA calculation.
Data values given in this report are set by consensus following discussions between
representatives from industry, the competent authorities and the central government. Data values
are often based on previously made decisions using best judgement of the available information at
that time. A number of chapters have been completed with an appendix, called Commentary, to
provide a record of the reasoning leading to specific data. These commentaries discuss
motivation for certain decisions and the base used for specific data and their validity.
Finally, please note that the information in this document should be used as a guideline to a QRA
calculation. The author of a QRA may deviate from the recommendations given here if site-
specific information demands it. However, deviations should be made in consultation with and be
approved by the competent authorities, with the motivation documented in the QRA report.
Selection of installations 2.1
2. SELECTION OF THE INSTALLATIONS FOR THE QRA
2.1 Introduction
A Quantitative Risk Assessment (QRA) is a valuable tool for determining the risk of the use,
handling, transport and storage of dangerous substances. QRAs are done if dangerous substances
are thought to be present at a location (e.g. industrial sites and transportation routes) in amounts
that can endanger the environment. A QRA is used in a Safety Report to demonstrate the risk
caused by the establishment and to provide the competent authority with relevant information
for assessing incremental risk and for enabling decisions on the acceptability of risk related to
developments on site of or around the establishment. A Safety Report should be made if the
amount of dangerous substances that can be present in an establishment exceeds a threshold value
[EU96]. The procedure to determine whether a Safety Report has to be made is given in the
Council Directive 96/82/EC of 9 December 1996 on the control of major-accident hazards
involving dangerous substances [EU96]
a
. The procedure is outlined in Appendix 2.A.
The total number of installations in an establishment where a Safety Report has to be made can
be very large. Since not all installations contribute significantly to the risk, it is not worthwhile to
include all installations in the QRA. Therefore a selection method, described below, has been
developed to indicate the installations that contribute most to the risk caused by the
establishment. These installations have to be considered in the QRA.
The method to select the installations to be included in the QRA is a general one and should
therefore be considered as a guidance only [NR]. Consequently, some installations can be
unjustly omitted. Notable examples are loading and unloading facilities, inter-unit pipelines, (by-)
products in the process, products formed through burning during a fire, combustion products and
reaction products from run-away reactions.
The installations to be considered in the QRA are selected following consultation between the
operator of an establishment and the competent authority. The operator of an establishment does
the calculations needed to select installations. However, the selection of installations is the
responsibility of the competent authority only. The competent authority can therefore decide to
include installations in the QRA that are not selected using the method described here.
The selection method applies only to the establishments for which a Safety Report has to be
made. If a QRA is made for transportation routes or other establishments, all installations have to
be included in the QRA. However, the competent authority may accept the application of the
selection method in these cases too.
In this chapter, the selection method is described. An example calculation is included in Appendix
2.B.
a
The reference should be replaced with the corresponding Dutch legislation when appropriate.
2.2 Exclusion of particular substances
Following Article 9, Paragraph 6 of Council Directive 96/82/EC on the control of major-accident
hazards involving dangerous substances [EU96], particular substances in a state incapable of
creating a major-accident hazard can be excluded from the Safety Report and, consequently, from
the QRA calculations. The decision is the responsibility of the competent authority. Criteria for
the decision that substances are in a state incapable of creating a major hazard accident are given
in the Commission decision on harmonized criteria for dispensations according to Article 9 of
Council Directive 96/82/EC of 9 December 1996 on the control of major-accident hazards
involving dangerous substances [EU98]. A particular dangerous substance can be excluded if at
least one of the following generic criteria is fulfilled.
1. Physical form of the substance
Substances in solid form, such that, under both normal conditions and any abnormal
conditions which can be reasonably foreseen, a release, of matter or of energy, which could
create a major-accident hazard, is not possible.
2. Containment and Quantities
Substances packaged or contained in such a fashion and in such quantities that the
maximum release possible under any circumstances cannot create a major-accident hazard.
3. Location and Quantities
Substances present in such quantities and at such distances from other dangerous
substances (at the establishment or elsewhere) that they can neither create a major-accident
hazard by themselves nor initiate a major accident involving other dangerous substances.
4. Classification
Substances which are defined as dangerous substances by virtue of their generic
classification in Annex I Part 2 of Council Directive 96/82/EC, but which cannot create a
major-accident hazard, and for which therefore the generic classification is inappropriate for
this purpose.
2.3 The selection method
If a Quantitative Risk Assessment has to be made as part of a Safety Report, it is not necessary
to assess the risks of all installations of an establishment. However, it is important to consider all
installations substantially contributing to the risk caused by an establishment. Therefore a
selection method, based on the amount of substance present in an installation and on the process
conditions, has been developed to determine which installations should be considered in a QRA.
The selection method consisting of the following steps is illustrated in
Figure 2.1 Outline of the selection method.
1. The establishment is divided into a number of independent installations, according to the
procedure given in Section 2.3.1.
2. For all other installations, the intrinsic hazard, induced from the amount of substance
present, the process conditions and the dangerous properties of the substance, is
determined. The indication number, A, shows the measure of intrinsic hazard of the
installation. This number is calculated according to the procedure given in Section 2.3.2.
3. The hazard of an installation is calculated for a number of points in the surroundings of the
establishment. The hazard at a point is induced knowing the indication number and the
distance between the point and the installation. The measure of the hazard at a given point
is indicated by the selection number, S, which is calculated according to the procedure given
in Section 2.3.3.
4. Installations are selected for analysis in the QRA on the basis of the relative magnitude of
the selection number according to the procedure outlined in Section 2.3.4.
2.3.1 Definition of installations in an establishment
The first step in the selection method is to divide an establishment into a number of separate
installations. This is a complex process, which may be open for discussion. This section offers
some guidance.
An important criterion for the definition of a separate installation is that loss of containment of
one installation does not lead to release of significant amounts of substances from other
installations. Consequently, two installations are considered separate if they can be isolated in a
very short time following an accident.
Two different types of installations are distinguished, i.e. process installations and storage
installations. A process installation can consist of several tanks, pipes and similar equipment. A
storage installation, like a storage tank, is always considered to be separate. Often a storage
installation is equipped with devices like recirculation systems and heat exchangers to keep the
substance at storage conditions. However, the installation is still considered as a storage
installation, whether or not such devices are present. The classification of transport units in an
establishment is described in Section 2.3.5.
Since the division into separate installations is a complex process, consultation between the
operator of the establishment and the competent authority is considered useful.
2.3.2 Calculation of the indication number, A
The intrinsic hazard of an installation depends on the amount of substance present, the physical
and toxic properties of the substance and the specific process conditions. The indication number,
A, is calculated as a measure of the intrinsic hazard of an installation.
The indication number, A, for an installation is a dimensionless number defined as:
A = Q O
1
O
2
O
3
(2.1)
G
with:
Q the quantity of substance present in the installation (kg), as described in Section 2.3.2.1.
O
i
the factors for process conditions (-), as described in Section 2.3.2.2.
G the limit value (kg), as described in Section 2.3.2.3.
2.3.2.1 Quantity of substance present, Q
The quantity of the substance present in an installation is the total amount of substance
contained within the installation, where the desired and undesired generation of substances in the
process should be considered, including a possible loss of control. The following rules apply:
Mixtures and preparations can be distinguished into two different types, i.e. (1) a
dangerous substance in a non-dangerous solvent and (2) a mixture of dangerous substances.
(1) If a dangerous substance is dissolved in a non-dangerous substance, only the amount
of dangerous substance has to be considered. Examples are ammonia in water or
hydrogen chloride in water. Mixtures and preparations of toxic substances need to be
taken into account in the selection process only if the mixture or preparation is
classified as (very) toxic.
(2) If a mixture of various dangerous substances has its own physical, chemical and toxic
properties, it should be treated in the same way as pure substances.
If dangerous substances are stored as small packaging units in one place and it is likely that
loss of containment occurs for a large number of packaging units simultaneously, the total
amount of the substance stored in that place has to be considered. Examples are the storage
of explosives or fireworks and the release of toxic combustion products during a fire.
For toxic substances in the solid state, only the amount of respirable powder has to be
considered. However, also the possibility of a fire has to be considered. A fire will result in
combustion products and an amount of unburned powder in the air.
Storage tanks can be used to store different substances at different times. If large numbers
of different substances are transshipped from an establishment, it is useful to classify the
substances and to use sample substances for each category in the QRA. A classification
method is described in [VVoW95]. It should be noted that if a specific substance
constitutes an important part of the total amount transshipped, the substance itself will
have to be considered.
2.3.2.2 Factors for process conditions, O
i
Three different factors are applied to account for process conditions:
O
1
a factor to account for process installation versus storage installation
O
2
a factor to account for the positioning of the installation
O
3
a factor to account for the amount of substance in the vapour phase after release, based on
the process temperature, the atmospheric boiling point, the substance phase and the
ambient temperature.
The factors for the process conditions apply to toxic and flammable substances only. For
explosives, O
1
= O
2
= O
3
= 1.
2.3.2.2.1 Factor O
1
Factor O
1
(see Table 2.1) accounts for the type of installation, be it for processing or storage.
Table 2.1 Factor O
1
to account for the type of installation
Type O
1
installation for processing 1
installation for storage 0.1
2.3.2.2.2 Factor O
2
Factor O
2
(see Table 2.2) accounts for the positioning of the installation and the presence of
provisions to prevent the substances disseminating into the environment.
Table 2.2 Factor O
2
to account for the positioning of the installation
Positioning O
2
outdoor installation 1.0
enclosed installation 0.1
installation situated in a bund and a process temperature, T
p
,
less than the atmospheric boiling point T
bp
plus 5 C, i.e. T
p
T
bp
+ 5 C
0.1
installation situated in a bund and a process temperature, T
p
,
more than the atmospheric boiling point T
bp
plus 5 C, i.e. T
p
> T
bp
+ 5 C
1.0
Notes:
1. For storage, the process temperature should be seen as the storage temperature.
2. The enclosure of the installation should prevent substances being spread in the
environment. This means that (a) the enclosure should remain unimpaired following the
physical pressures due to the instantaneous release of the installation inventory and (b) the
enclosure should reduce significantly the direct release into the atmosphere. A guideline: if
the enclosure reduces the source term into the atmosphere by more than a factor 5, or if the
enclosure redirects the release to a safe outlet, the installation will be considered enclosed,
otherwise it is an outdoor installation.
3. A bund should prevent the substance spreading in the environment.
4. A second containment designed to contain the liquid and withstand all possible loads is
interpreted as a bund, O
2
= 0.1. Factor 0.1 applies to double containment atmospheric
tanks, full containment atmospheric tanks, in-ground atmospheric tanks and mounded
atmospheric tanks.
2.3.2.2.3 Factor O
3
Factor O
3
(see Table 2.3) accounts for the process conditions and is a measure of the amount of
substance in the gas phase after its release.
Table 2.3 Factor O
3
to account for the process conditions
Phase O
3
substance in gas phase 10
substance in liquid phase
- saturation pressure at process temperature of 3 bar or higher 10
- saturation pressure at process temperature of between 1 to 3 bar X +
- saturation pressure at process temperature of less than 1 bar P
i
+
substance in solid phase 0.1
Notes:
1. For storage, the process temperature should be seen as the storage temperature..
2. Pressures are absolute.
3. Factor X increases linearly from 1 to 10 as the saturation pressure at process temperature,
P
sat
, increases from 1 to 3 bar. In the equation, where P
sat
is given in bars:
X = 4.5 P
sat
3.5 (2.2)
4. P
i
is equal to the partial vapour pressure (in bars) of the substance at process temperature.
5. If the substance is in the liquid phase, an amount, , is added to account for the extra
evaporation due to the heat flux from the environment to the liquid pool formed. The value
of (see Table 2.4) depends only on the atmospheric boiling point, T
bp
Table 2.4 Added value accounts for liquid pool evaporation
25 C T
bp
0
75 C T
bp
< 25 C 1
125 C T
bp
< 75 C 2
T
bp
< 125 C 3
A 10% point should be used for mixtures of dangerous substances, i.e. the temperature at
which 10% of the mixture is distilled off.
6. For dangerous substances in non-dangerous solvents, the partial vapour pressure of the
dangerous substance at process temperature is to be used for the saturation pressure at
process temperature. The factor X increases linearly from 1 to 10 if the partial vapour
pressure of the dangerous substance at process temperature increases from 1 to 3 bar.
7. The factor O
3
is limited to a minimum value of 0.1 and a maximum value of 10.
2.3.2.3 Limit value, G
The limit value, G, is a measure of the dangerous properties of the substance based on both the
physical properties and the toxic/explosive/flammable properties of the substance.
2.3.2.3.1 Limit value for toxic substances
The limit value for toxic substances (see Table 2.5) is determined by the lethal concentration,
LC
50
(rat, inh, 1h) and the phase at 25 C.
Notes:
1. The phase of the substance (gas, liquid and solid) assumes a temperature of 25 C. In
addition, the following subdivision holds for liquids:
Liquid (L) atmospheric boiling point T
bp
between 25 C and 50 C
Liquid (M) atmospheric boiling point T
bp
between 50 C and 100 C
Liquid (H) atmospheric boiling point T
bp
above 100 C
2. LC
50
(rat, inh, 1h) is the LC
50
value for rats using an inhalation method for exposure of one
hour. These values are listed for a number of toxic substances in the database [RIVM99].
3. The limit value should be derived from Table 2.5. Limit values to determine the Report on
Occupational Safety compliance are listed for a number of substances in [SZW97] and
[RIVM99]. These values can differ from those derived from Table 2.5 for some
carcinogenic substances and if new toxicity data are used.
Table 2.5 Limit value, G, for toxic substances
LC
50
(rat, inh, 1h) (mg m
-3
) Phase at 25 C Limit value (kg)
LC 100 gas 3
liquid (L) 10
liquid (M) 30
liquid (H) 100
solid 300
100 < LC 500 gas 30
liquid (L) 100
liquid (M) 300
liquid (H) 1000
solid 3000
500 < LC 2000 gas 300
liquid (L) 1000
liquid (M) 3000
liquid (H) 10,000
solid
2000 < LC 20,000 gas 3000
liquid (L) 10,000
liquid (M)
liquid (H)
solid
LC > 20,000 all phases
2.3.2.4 Limit value for flammable substances
The limit value for flammables is 10,000 kg.
Note:
1. Flammables are defined for the selection system as substances having a process
temperature equal to or higher than the flashpoint. The flashpoint is determined using the
apparatus of Abel-Pensky for flame points up to and including 65
0
C and the apparatus of
Pensky-Martens for flame points higher than 65
0
C.
2.3.2.5 Limit value for explosive substances
The limit value for explosive substances is the amount of substance (in kg) which releases an
amount of energy equivalent to 1000 kg TNT (explosion energy 4600 kJ/kg).
2.3.2.6 Calculation of the indication number
The indication number, A
i
, of an installation for a substance i is calculated as:
A
i
= Q
1
O
1
O
2
O
3
(2.3)
G
i
with:
Q
i
the quantity of substance i present in the installation (in kg)
O
1
the factor for installation type, whether process or storage (-)
O
2
the factor for the positioning of the installation, enclosed, bund or outdoors (-)
O
3
the factor for the process conditions (-)
G
i
the limit value of substance i (in kg).
For explosives, O
1
= O
2
= O
3
= 1 and, consequently, A = Q / G.
Various substances and process conditions can be present within one installation. In this case, an
indication number, A
i,p
, is calculated for every substance, i, and for every process condition, p.
The indication number, A, for an installation is calculated as the sum over all indication numbers,
i,p
A
i,p
. This sum is calculated for three different groups of substances separately, namely,
flammables (A
F
), toxics (A
T
) and explosives (A
E
).
A
T
=
i,p
A
i,p
, sum over all toxic substances and process conditions
A
F
=
i,p
A
i,p
, sum over all flammable substances and process conditions
A
E
=
i,p
A
i,p
, sum over all explosive substances and process conditions
An installation can have up to three different indication numbers.
Note:
1. If a substance belongs to more than one group, an indication number is calculated for each
group separately. For instance, if a substance is both toxic and flammable, two indication
numbers, A
i,p
, are calculated:
A
T
i,p
for the substance as a toxic using the total quantity, Q
i
, and the limit value, G
T
i
,
corresponding with the toxic properties.
A
F
i,p
for the substance as a flammable using the total quantity, Q
i
, and the corresponding
limit value for flammables, G
F
i
= 10,000 kg.
2.3.3 Calculation of the selection number, S
The selection number, S, is a measure of the hazard of an installation at a specific location and is
calculated by multiplying the indication number, A, of an installation by a factor (100/L)
2
for
toxic substances and a factor (100/L)
3
for flammable or explosive substances. Again, three
different selection numbers can exist for one installation:
for toxics (2.4)
for flammables (2.5)
for explosives (2.6)
L is the distance from the installation to the specific location in metres, with a minimum of
100 m.
The selection number has to be calculated for every installation at a minimum of eight locations
on the boundary of the establishment. The distance between two adjacent locations must not be
more than 50 metres. The selection number must be calculated for the total boundary of the
establishment, even if the establishment borders on a similar establishment. If the establishment
is bounded by surface water, the selection number must be calculated on the bank side situated
opposite the establishment.
Besides the calculation on the boundary of the establishment, the selection number, S, must also
be calculated for every installation at a location in a residential area, existing or planned, closest to
the installation.
2.3.4 Selection of installations
An installation is selected for analysis in a QRA if:
the selection number of an installation is larger than one at a location on the boundary of the
establishment (or on the bank side situated opposite the establishment) and larger than
50% of the maximum selection number at that location.
or
the selection number of an installation is larger than one at a location in the residential area,
existing or planned, closest to the installation.
Note:
1. The effects of the release of toxic substances may extend further than the effects of the
release of flammable substances. If only installations with flammable substances are
selected and the selection number of an installation with a toxic substance is in the same
order of magnitude as the maximum selection number, the installation with toxic substances
should also be included.
2.3.5 Specific problems
2.3.5.1 Inter-unit pipelines
Large inter-unit pipelines in an establishment can contribute considerably to the risk caused by
the establishment e.g.:
inter-unit pipelines may be situated near the boundary of an establishment,
inter-unit pipelines may release large amounts of substances due to their own hold-up and
the feed from the upstream vessel, and
inter-unit pipelines may have large failure frequencies.
For the selection method, the quantity present is calculated as:
For pipelines containing liquids or pure gases, the quantity present is set equal to the
amount in the pipeline, with a length equal to 600 seconds multiplied by the velocity of the
liquid or gas in the pipeline.
For pipelines containing liquefied pressurized gases, the quantity present is a function of
the diameter of the pipeline and the substance. The quantity present is equal to the amount
present in a pipeline, with a length that is emptied after 600 seconds. For a number of
reference substances, the length of the pipeline emptied is given in Figure 1.2. For all other
substances, the length can be estimated using the physical properties of the substance,
particularly vapour pressure at 10 C, to select one of the curves in Figure 1.2.
If the length of the pipeline calculated exceeds the actual length of the pipeline, the quantity
present is equal to the amount between two quick-closing blocking valves isolating the pipeline at
an incident. The time needed to close the two blocking valves is assumed to be so short that the
amount released when the blocking valves are open is small compared to the amount between the
two blocking valves. If not, the amount between the two blocking valves should be corrected with
the mass released during the time the blocking valves are open. However, the quantity present
should not exceed the amount in the length of the pipeline equal to 600 seconds multiplied by the
velocity of the liquid or gas, or the length of the pipeline emptied after 600 sec (liquefied
pressurized gases).
The factors for the process conditions O
1
- O
3
apply. An inter-unit pipeline should be considered
as a process installation, O
1
= 1. The factors O
2
and O
3
are given in Table 2.2 and Table 2.3. An
underground inter-unit pipeline is to be considered enclosed (O
2
= 0.1).
To calculate the selection number, various points on the pipeline should be considered for the
location of the total quantity present. The distance between two neighbouring points must be
equal to circa 50 metres.
To select pipelines for a QRA calculation, a division is made in pipelines included in the
establishments permit and pipelines not included. If an inter-unit pipeline is included in the
permit, the pipeline should be dealt with like all other installations. However, if an inter-unit
pipeline is not included in the permit, the installations without these inter-unit pipelines are first
selected. This results in a list of installations in the establishment. Next, a new selection is made
to include the inter-unit pipelines not included in the permit. This results in an additional list of
inter-unit pipelines to be considered in the QRA.
If an inter-unit pipeline is selected on the basis of the selection number of one or more release
locations, the total inter-unit pipeline will have to be included in the QRA.
Figure 2.2 Length of pipeline emptied after 600 s for a number of reference substances for a two
phase outflow at 10 C.
2.3.5.2 Loading and unloading activities
During loading and unloading activities, storage tanks are situated within the transport unit at the
establishment. Three installations have to be considered for the selection procedure, namely, the
storage tank in the transport unit, the loading facility and the connecting installation at the
establishment. The following rules apply:
the storage tank in the transport unit is considered a process installation if the time that
the transport unit is connected to a process installation is less than one day. In all other
cases, the storage tank in the transport unit is considered to be a storage installation.
the loading facility is a process installation and should be included in the QRA if either the
supplying or the receiving installation is selected.
Storage tanks on ships should be included if the presence of the ship is connected to the
establishment. Only the substances involved in loading and unloading activities have to be
considered for the selection. If a storage tank on a ship is to be considered, installations
without the storage tank on the ship are first selected. This results in a list of installations
of the establishment. Next, a new selection is made of installations with the storage tank on
the ships included. This results in an additional list of installations for consideration in the
QRA.
Transport units are only present part of the time. Although this is important in the QRA,
it is not considered in the selection procedure.
Appendix 2.A Procedure to assess the obligation to make a Safety Report
2.A.1 Outline of the procedure
The procedure, outlined below, to determine whether a Safety Report should be obligatory was
taken up in the Council Directive 96/82/EC of 9 December 1996 on the control of major-accident
hazards involving dangerous substances [EU96]. It should be noted that the outline given here is
only a short description of the framework and should not be seen as the complete procedure. The
rules and notes in Annex I of the Council Directive 96/82/EC of 9 December 1996, taken up in
2.A.2, are decisive and should be considered carefully.
The procedure:
1. Determine the substances present in the establishment. Presence of substances is taken to
mean the actual or anticipated presence of such substances in the establishment, or the
presence of those substances believed to be possibly generated during a chemical process
which has got out of control.
Notes:
- If a substance is licensed, it is assumed to be present.
- The presence of a substance is meant to refer to a substance in the establishment for at
least five consecutive days or at a frequency of more than 10 times per year.
2. Determine for each substance, x, the maximum quantity present or likely to be present at
any one time, q
x
.
Note:
- The licensed amount of substance is assumed to be present.
3. Look for the substance x in the table of Part 1 of Annex I
If substance x is named in the table of Part 1, determine the corresponding qualifying
quantity, Q
x
, in column 3 (Article 9).
If substance x is not named in the table of Part 1, determine in which category of the
table the substance falls. Determine the corresponding qualifying quantity, Q
x
, in
column 3 (Article 9).
4. Determine for each substance x the value q
x
/ Q
x
. If q
x
/ Q
x
> 1 for one or more of the
substances, a Safety Report should be made.
5. If q
x
/ Q
x
< 1 for all substances x, the sum q
1
/ Q
1
+ q
2
/ Q
2
+q
3
/ Q
3
+ .... has to be
calculated for two groups of substances separately, namely, for all substances classified in
the categories 1, 2 and 9, and for all substances classified in the categories 3, 4, 5, 6, 7a, 7b,
and 8. If one of the two sums is larger than 1, a Safety Report should be made.
Named substances should be classified and added accordingly to the categories in the table
of Part 2, using the qualifying quantity Q
x
of the table in Part 1.
There are databases available which give the classification of a number of dangerous substances,
e.g. the database of substances of RIVM [RIVM99].
2.A.2 Annex I of the Council Directive 96/82/EC of 9 December 1996
APPLICATION OF THE DIRECTIVE
INTRODUCTION
1. This Annex applies to the presence of dangerous substances at any establishment within the meaning of
Article 3 of this Directive and determines the application of the relevant Articles thereof.
2. Mixtures and preparations shall be treated in the same way as pure substances provided they remain within
concentration limits set according to their properties under the relevant Directives given in Part 2, Note 1, or
their latest adaptation to technical progress, unless a percentage composition or other description is
specifically given.
3. The qualifying quantities set out below relate to each establishment.
4. The quantities to be considered for the application of the relevant Articles are the maximum quantities which
are present or are likely to be present at any one time. Dangerous substances present at an establishment only
in quantities equal to or less than 2 % of the relevant qualifying quantity shall be ignored for the purposes of
calculating the total quantity present if their location within an establishment is such that it cannot act as an
initiator of a major accident elsewhere on the site.
5. The rules given in Part 2, Note 4, governing the addition of dangerous substances, or categories of dangerous
substances, shall apply where appropriate.
PART 1
Named substances
Where a substance or group of substances listed in Part 1 also falls within a category of Part 2, the qualifying
quantities set out in Part 1 must be used.
Column 1 Column 2 Column 3
Qualifying quantity (tonnes)
Dangerous substances for the application of
Articles 6 and 7 Article 9
Ammonium nitrate 350 2500
Ammonium nitrate 1250 5000
Arsenic pentoxide, arsenic (V) acid and/or salts 1 2
Arsenic trioxide, arsenious (III) acid and/or salts 0,1
Bromine 20 100
Chlorine 10 25
Nickel compounds in inhalable powder form (nickel
monoxide, nickel dioxide, nickel sulphide, trinickel
disulphide, dinickel trioxide)
1
Ethyleneimine 10 20
Fluorine 10 20
Formaldehyde (concentration 90 %) 5 50
Hydrogen 5 50
Hydrogen chloride (liquefied gas) 25 250
Lead alkyls 5 50
Dangerous substances for the application of
Liquefied extremely flammable gases (including LPG)
and natural gas
50 200
Acetylene 5 50
Ethylene oxide 5 50
Propylene oxide 5 50
Methanol 500 5000
4, 4-Methylenebis (2-chloraniline) and/or salts, in
powder form
0.01
Methylisocyanate 0.15
Oxygen 200 2000
Toluene diisocyanate 10 100
Carbonyl dichloride (phosgene) 0.3 0.75
Arsenic trihydride (arsine) 0.2 1
Phosphorus trihydride (phosphine) 0.2 1
Sulphur dichloride 1 1
Sulphur trioxide 15 75
Polychlorodibenzofurans and
polychlorodibenzodioxins (including TCDD),
calculated in TCDD equivalent
0,001
The following CARCINOGENS:
4-Aminobiphenyl and/or its salts,
Benzidine and/or salts,
Bis (chloromethyl) ether,
Chloromethyl methyl ether,
Dimethylcarbamoyl chloride,
Dimethylnitrosomine,
Hexamethylphosphoric triamide,
2-Naphtylamine and/or salts,
and 1,3 Propanesultone 4-nitrodiphenyl
0,001 0.001
Automotive petrol and other petroleum spirits 5000 50,000
NOTES
1. Ammonium nitrate (350 / 2500)
This applies to ammonium nitrate and ammonium nitrate compounds in which the nitrogen content as a result
of the ammonium nitrate is more than 28 % by weight (compounds other than those referred to in Note 2) and to
aqueous ammonium nitrate solutions in which the concentration of ammonium nitrate is more than 90 % by
weight.
2. Ammonium nitrate (1250/5000)
This applies to simple ammonium-nitrate based fertilizers which comply with Directive 80/876/EEC and to
composite fertilizers in which the nitrogen content as a result of the ammonium nitrate is more than 28% in
weight (a composite fertilizer contains ammonium nitrate with phosphate and/or potash).
3. Polychlorodibenzofurans and polychlorodibenzodioxins
The quantities of polychlorodibenzofurans and polychlorodibenzodioxins are calculated using the following
factors:
International Toxic Equivalent Factors (ITEF) for the congeners of concern (NATO/CCMS)
2,3,7,8-TCDD 1 2,3,7,8-TCDF 0.1
1,2,3,7,8-PeDD 0.5 2,3,4,7,8-PeCDF 0.5
1,2,3,7,8-PeCDF 0.05
1,2,3,4,7,8-HxCDD 0.1
1,2,3,6,7,8-HxCDD 0.1 1,2,3,4,7,8-HxCDF 0.1
1,2,3,7,8,9-HxCDD 0.1 1,2,3,7,8,9-HxCDF 0.1
1,2,3,6,7,8-HxCDF 0.1
1,2,3,4,6,7,8-HpCDD 0.01 2,3,4,6,7,8-HxCDF 0.1
OCDD 0.001 1,2,3,4,6,7,8-HpCDF 0.1
1,2,3,4,7,8,9-HpCDF 0.01
OCDF 0.01
(T = tetra, P = penta, Hx = hexa, Hp = hepta, O = octa)
PART 2
Categories of substances and preparations not specifically named in Part 1
of dangerous substances
Categories of dangerous substances as delivered in Article 3 (4),
for the application of
1. VERY TOXIC 5 20
2. TOXIC 50 200
3. OXIDIZING 50 200
4. EXPLOSIVE (where the substance or
preparation falls within the definition given
in Note 2 (a))
50 200
5. EXPLOSIVE (where the substance or
in Note 2 (b))
10 50
6. FLAMMABLE (where the substance or
in Note 3 (a))
5000 50,000
7 a. HIGHLY FLAMMABLE (where the
substance or preparation falls within the
definition given in Note 3 (b) (1)) 50 200
7
b.
HIGHLY FLAMMABLE liquids (where the
definition given in Note 3 (b) (2)) 5000 50,000
8. EXTREMELY FLAMMABLE (where the
definition given in Note 3 (c))
10 50
9. DANGEROUS FOR THE ENVIRONMENT
in combination with risk phrases:
(i) R50: Very toxic to aquatic organisms
(ii) R51:Toxic to aquatic organisms; and
R53: May cause long term adverse effects
in the aquatic environment
200
500
500
2000
10. ANY CLASSIFICATION not covered by
those given above in combination with risk
phrases:
(i) R14: Reacts violently with water
(including R14/15)
(ii) R29: in contact with water, liberates
toxic gas
100
50
500
200
NOTES
1. Substances and preparations are classified according to the following Directives (as amended) and their
current adaptation to technical progress:
Council Directive 67/548/EEC of 27 June 1967 on the approximation of the laws, regulations and
administrative provisions relating to the classification, packaging and labelling of dangerous substances (
b
),
Council Directive 88/379/EEC of 7 June 1988 on the approximation of the laws, regulations and
administrative provisions of the Member States relating to the classification, packaging and labelling of
dangerous preparations (
c
),
Council Directive 78/631/EEC of 26 June 1978 on the approximation of the laws of the Member States
relating to the classification, packaging and labelling of dangerous preparations (pesticides) (
d
).
In the case of substances and preparations which are not classified as dangerous according to any of the above
Directives but which nevertheless are present, or are likely to be present, in an establishment and which possess
or are likely to possess, under the conditions found at the establishment, equivalent properties in terms of
major-accident potential, the procedures for provisional classification shall be followed according to the relevant
Article of the appropriate Directive.
In the case of substances and preparations with properties giving rise to more than one classification , for the
purposes of this Directive the lowest thresholds shall apply.
For the purposes of this Directive, a list providing information on substances and preparations shall be
established, kept up to date and approved by the procedure set up under Article 22.
2. An explosive means:
(a) (i) a substance or preparation which creates the risk of an explosion by shock, friction, fire or other
sources of ignition (risk phrase R 2),
(ii) a pyrotechnic substance is a substance (or mixture of substances) designated to produce heat, light,
sound, gas or smoke or a combination of such effects through non-detonating self-sustained
exothermic chemical reactions, or
(iii)an explosive or pyrotechnic substance or preparation contained in objects;
(b) a substance or preparation which creates extreme risks of explosion by shock, friction, fire or other
sources of ignition (risk phrase R 3).
3. Flammable, highly flammable, and extremely flammable in categories 6, 7 and 8 mean:
(a) flammable liquids:
substances and preparations having a flash point equal to or greater than 21 C and less than or equal to
55C (risk phrase R 10), supporting combustion;
(b) highly flammable liquids:
1. substances and preparations which may become hot and finally catch fire in contact with air at
ambient temperature without any input of energy (risk phrase R 17),
substances which have a flash point lower than 55 C and which remain liquid under pressure,
where particular processing conditions, such as high pressure or high temperature, may create
major-accident hazards;
2. substances and preparations having a flash point lower than 21 C and which are not extremely
flammable (risk phrase R 11, second indent);
(
b
) OJ No. 196, 16. 8. 1967, p. l. Directive as fast <OK?> amended by Directive 93/105/EC (OJ No. L 294, 30.
11. 1993, p. 2l).
(
c
) OJ No. L 187, 16. 7. 1988, p. 14.
(
d
) OJ No. L 206, 29. 7. 1978, p. 13. Directive as fast amended by Directive 92/32/EEC (OJ No. L 154, 5. 6.
1992, p. 1).
(c) extremely flammable gases and liquids:
1. liquid substances and preparations which have a flash point lower than 0 C and the boiling point (or,
in the case of a boiling range, the initial boiling point) of which at normal pressure is less than or
equal to 35 C (risk phrase R 12, first indent), and
2. gaseous substances and preparations which are flammable in contact with air at ambient temperature
and pressure (risk phrase R 12, second indent), whether or not kept in the gaseous or liquid state
under pressure, excluding liquefied extremely flammable gases (including LPG) and natural gas
referred to in Part 1, and
3. liquid substances and preparations maintained at a temperature above their boiling point.
4. The addition of dangerous substances to determine the quantity present at an establishment shall be carried
out according to the following rule:
if the sum
q1/Q + q2/Q + q3/Q + q4/Q + q5/Q + ... 1,
where qx = the quantity of dangerous substances x (or category of dangerous substances) falling within Parts
1 or 2 of this Annex,
Q = the relevant threshold quantity from Parts 1 or 2,
then the establishment is covered by the relevant requirements of this Directive.
This rule will apply for the following circumstances:
(a) for substances and preparations appearing in Part 1 at quantities less than their individual qualifying
quantity present with substances having the same classification from Part 2, and the addition of
substances and preparations with the same classification from Part 2;
(b) for the addition of categories 1, 2 and 9 present at an establishment together;
(c) for the addition of categories 3, 4, 5, 6, 7a, 7b and 8, present at an establishment together.
Appendix 2.B An example calculation
2.B.1 Description of the establishment and the installations
An establishment contains five separate installations. The area of the establishment is rectangular
between the lower left point ( 400 m, 200 m) and the upper right point (+300 m, +300 m). A
residential area is situated to the north of the establishment, at 400 m from its centre.
The installations, I
i
, are listed in Table 2.B.1.
Table 2.B.1 Installations, I
i
, present at the establishment
No Location Process
I
1
(200, 200) Production installation inside a building, containing pure chlorine in an
amount of 2100 kg at a process temperature of 35 C (vapour pressure at a
process temperature of 10 bar)
I
2
(0, 0) Production installation outdoors. The installation contains various
flammable substances at different process conditions:
ethylene amount 200,000 kg liquid at 30 C (vapour pressure 20 bar)
ethane amount 100,000 kg gas at 80 C
butane amount 10,000 kggas at 30 C
propylene amount 10,000 kgliquid at 35 C (vapour pressure 1.75 bar)
propane amount 50,000 kgliquid at 80 C (vapour pressure 31 bar)
I
3
( 300, 150) Installation for storage of a 30% solution of hydrochloric acid in water. The
storage tank is situated outdoors and contains 1,500,000 kg solution at a
temperature of 25 C (partial vapour pressure P
i
= 0.02 bar).
I
4
(200, 100) The storage tank is connected to a process installation inside a building
where an amount of 300,000 kg of the 30% solution of hydrochloric acid in
water is processed at a temperature of 100 C (liquid, partial vapour
pressure of P
i
= 1.1 bar).
I
5
( 300, 125) A process installation outdoors contains pure ammonia (gas, 12,000 kg), a
60% solution of ammonia in water (9000 kg solution at 43 C, with a partia
vapour pressure P
i
= 9.4 bar). In the installation petrol is used (1000 kg) at
a temperature of 150 C.
The layout of the plant and the residential area is shown in Figure 2.B.1.
Figure 2.B.1 Layout of the plant and residential area, showing points closest to the installation
(A-C). Indicated are the locations (solid circle) of the installations (I.1-I.5) and the
locations (solid square) where the selection numbers are calculated. The locations
1, 2, 3, .. and A-C correspond with the points in Table 2.B.5.
2.B.2 Calculation of the indication number
2.B.2.1 Installation I
1
Installation I
1
is a process installation (O
1
= 1) situated in a building (O
2
= 0.1). One substance,
chlorine, is present in a quantity Q of 2100 kg. As the vapour pressure of chlorine is more than 3
bar, O
3
= 10. Chlorine is a toxic substance; in the gas phase at 25 C; LC
50
(rat, ihl, 1hr) = 293
ppm [SZW97]. The limit value is equal to G = 300 kg. Therefore, A
T
1
= 7.
2
Installation I
2
1
= 1) situated outdoors (O
2
= 1). Five different
combinations of substances and process conditions are present, as shown in Table 2.B.2.
Table 2.B.2 Combinations of substances and process conditions present at Installation I
2
Substance Q O
3
G A
F
Note
ethylene 200,000 kg 10 10,000 kg 200 1
ethane 100,000 kg 10 10,000 kg 100 2
butane 10,000 kg 10 10,000 kg 10 3
propylene 10,000 kg 5.4 10,000 kg 5.4 4
propane 50,000 kg 10 10,000 kg 50 5
Notes:
1. Ethylene is a flammable substance having a vapour pressure greater than 3 bar under the
process conditions.
2. Ethane is a flammable substance in the gas phase under the process conditions.
3. Butane is a flammable substance in the gas phase under the process conditions.
4. Propylene is a flammable substance. The vapour pressure of propylene, P
i
, is equal to
1.75 bar at the process temperature, T
p
= 35 C. Therefore X = 4.5 1.75 3.5 = 4.4.
The boiling point, T
bp
, is equal to 48 C. Therefore = 1 and O
3
= 5.4.
5. Propane is a flammable substance having a vapour pressure greater than 3 bar under the
process conditions.
3
Installation I
3
is for storage (O
1
= 0.1) and situated outdoors (O
2
= 1). The amount of hydrogen
chloride present is 30% of 1,500,000 kg solution; Q = 450,000 kg. The substance, 30% solution
of hydrochloric acid in water, is a liquid. The partial vapour pressure of the dangerous substance,
hydrogen chloride, is P
i
= 0.02 bar; therefore X = 0.02. The boiling point of the substance, 30%
solution of hydrochloric acid in water, is 57 C, so = 0. As the resulting O
3
is less than the
minimum value, 0.1, O
3
= 0.1. Hydrogen chloride is a toxic substance and in the gas phase at 25
C; LC
50
(rat, ihl, 1hr) = 3124 ppm [SZW97]. The limit value is equal to G = 3000 kg. Therefore
A
T
3
= 1.5.
4
Installation I
4
1
= 1) and situated inside a building (O
2
= 0.1). The
amount of hydrogen chloride present is 30% of 300,000 kg solution; Q = 90,000 kg. The partial
vapor pressure of hydrochloric acid is P
i
= 1.1 bar at T
p
= 100 C. The factor X = 4.5 1.1 3.5
= 1.5. The boiling point of the substance, 30% solution of hydrochloric acid in water, is 57 C,
so = 0 and O
3
= 1.5. The limit value is equal to G = 3000 kg. Therefore A
T
4
= 4.5.
5
Installation I
5
1
= 1) and situated outdoors (O
2
= 1). Three
combinations of substances and process conditions are present. Furthermore, ammonia is both
toxic and flammable; both these hazards should be considered. The combinations of substances
and process conditions are shown in Table 2.B.3.
Table 2.B.3 Combinations of substances and process conditions present at Installation I
5
Substance Q O
3
G A
F
A
T
Note
ammonia, pure 12,000 kg 10 3,000 kg 40 1
ammonia, pure 12,000 kg 10 10,000 kg 12 1
ammonia, solution 5400 kg 10 3,000 kg 18 2
ammonia, pure 5400 kg 10 10,000 kg 5.4 2
petrol 1000 kg 10 10,000 kg 1 3
Notes:
1. Ammonia is a gas under the process conditions. The Limit value for the toxic substance
ammonia is equal to 3000 kg since ammonia is a gas at 25 C and LC
50
(rat, inh., 1hr) =
11,590 mg m
-3
[SZW97]. The limit value for the flammable substance ammonia is equal to
10,000 kg.
2. The quantity of ammonia present in solution is equal to 60% of 9000 kg solution, Q =
5400 kg. As the partial vapour pressure exceeds 3 bar, then O
3
= 10. The limit value for the
toxic substance ammonia is equal to 3000 kg since ammonia is a gas at 25 C and LC
50
(rat,
inh., 1hr) = 11,590 mg m
-3
[SZW97]. The limit value for the flammable substance ammonia
is equal to 10,000 kg.
3. Petrol is a flammable substance. The process temperature is higher than the 10% point. The
vapour pressure at 150 C has to be determined. For the example we assume it to be greater
than 3 bar. Therefore O
3
= 10.
2.B.2.6 Summary
The result of calculating the indication number is summarised in Table 2.B.4.
The indication numbers are:
installation I
1
A
T
= 7
installation I
2
A
F
= 365
installation I
3
A
T
= 1.5
installation I
4
A
T
= 4.5
installation I
5
A
T
= 58, A
F
= 18.4
Table 2.B.4 Indication numbers of the installations
Inst. Substance Type O
1
O
2
O
3
Q G A
i
I
1
chlorine T 1 0.1 10 2100 kg 300 kg 7
I
2
ethylene F 1 1 10 200,000 kg 10,000 kg 200
ethane F 1 1 10 100,000 kg 10,000 kg 100
butane F 1 1 10 10,000 kg 10,000 kg 10
propylene F 1 1 5.4 10,000 kg 10,000 kg 5.4
propane F 1 1 10 50,000 kg 10,000 kg 50
I
3
30%-HCl T 0.1 1 0.1 450,000 kg 3000 kg 1.5
I
4
30%-HCl T 1 0.1 1.5 90,000 kg 3000 kg 4.5
I
5
ammonia (g) T 1 1 10 12,000 kg 3000 kg 40
ammonia (s) T 1 1 10 5400 kg 3000 kg 18
ammonia (g) F 1 1 10 12,000 kg 10,000 kg 12
ammonia (s) F 1 1 10 5400 kg 10,000 kg 5.4
petrol F 1 1 10 1000 kg 10,000 kg 1
2.B.3 Calculation of the selection number
The selection number has to be calculated for points on the site boundary and residential area.
There are 48 points selected at 50-m intervals along the boundary (see Figure ). Furthermore, for
each installation the point in the plant area closest to the installation is selected. The selection
number is calculated from the distance of each point to the installation (minimal 100 metres). The
results are shown in Table 2.B.5. Installations 1, 2 and 5 have been selected for a QRA.
Table 2.B.5 Selection numbers at the positions selected
No. x y S1 S2 S3 S4 S5
T
S5
F
Selected
1 25 300 1.7 13.4 0.0 0.6 2.0 0.1 2
2 75 300 2.7 12.3 0.0 0.8 1.8 0.1 2
3 125 300 4.5 10.6 0.0 1.0 1.6 0.1 2
4 175 300 6.6 8.7 0.0 1.1 1.4 0.1 1, 2
5 225 300 6.6 6.9 0.0 1.1 1.3 0.1 1, 2
6 275 300 4.5 5.4 0.0 1.0 1.1 0.0 1, 2
7 300 275 4.5 5.4 0.0 1.1 1.1 0.0 1, 2
8 300 225 6.6 6.9 0.0 1.8 1.2 0.1 1, 2
9 300 175 6.6 8.7 0.0 2.9 1.3 0.1 1, 2
10 300 125 4.5 10.6 0.0 4.2 1.4 0.1 2
11 300 75 2.7 12.3 0.0 4.2 1.5 0.1 2
12 300 25 1.7 13.4 0.0 2.9 1.5 0.1 2
13 300 25 1.2 13.4 0.0 1.8 1.6 0.1 2
14 300 75 0.8 12.3 0.0 1.1 1.6 0.1 2
15 300 125 0.6 10.6 0.0 0.7 1.6 0.1 2
16 300 175 0.5 8.7 0.0 0.5 1.6 0.1 2
17 275 200 0.4 9.3 0.0 0.5 1.7 0.1 2
18 225 200 0.4 13.4 0.1 0.5 2.1 0.1 2
19 175 200 0.4 19.4 0.1 0.5 2.5 0.2 2
20 125 200 0.4 27.8 0.1 0.5 3.1 0.2 2
21 75 200 0.4 37.5 0.1 0.4 4.0 0.3 2
22 25 200 0.4 44.6 0.1 0.4 5.2 0.5 2
23 25 200 0.3 44.6 0.2 0.3 7.1 0.8 2
24 75 200 0.3 37.5 0.3 0.3 10.3 1.3 2
25 125 200 0.3 27.8 0.5 0.2 16.0 2.6 2, 5
26 175 200 0.2 19.4 0.8 0.2 27.3 5.8 2, 5
27 225 200 0.2 13.4 1.5 0.2 51.6 15.1 5
28 275 200 0.2 9.3 1.5 0.1 58.0 18.0 5
29 325 200 0.2 6.6 1.5 0.1 58.0 18.0 5
30 375 200 0.1 4.8 1.5 0.1 51.6 15.1 5
31 400 175 0.1 4.4 1.4 0.1 46.4 12.9 5
32 400 125 0.2 5.0 1.4 0.1 58.0 18.0 5
33 400 75 0.2 5.4 1.0 0.1 46.4 12.9 5
34 400 25 0.2 5.7 0.6 0.1 29.0 6.4 5
35 400 25 0.2 5.7 0.4 0.1 17.8 3.1 5
36 400 75 0.2 5.4 0.2 0.1 11.6 1.6 5
37 400 125 0.2 5.0 0.2 0.1 8.0 0.9 2, 5
38 400 175 0.2 4.4 0.1 0.1 5.8 0.6 2, 5
39 400 225 0.2 3.8 0.1 0.1 4.4 0.4 2, 5
40 400 275 0.2 3.2 0.1 0.1 3.4 0.3 2, 5
41 375 300 0.2 3.3 0.1 0.1 3.1 0.2 2, 5
42 325 300 0.2 4.2 0.1 0.1 3.2 0.2 2, 5
43 275 300 0.3 5.4 0.1 0.2 3.2 0.2 2, 5
44 225 300 0.4 6.9 0.1 0.2 3.1 0.2 2
45 175 300 0.5 8.7 0.1 0.2 3.0 0.2 2
46 125 300 0.6 10.6 0.1 0.3 2.7 0.2 2
47 75 300 0.8 12.3 0.1 0.4 2.5 0.2 2
48 25 300 1.2 13.4 0.1 0.5 2.3 0.1 2
C 200 400 1.8 1
B 0 400 5.7 2
A 300 400 0.0
C 200 400 0.5
A 300 400 2.1 0.12 5
Appendix 2.C Commentary
The procedure to select installations for the QRA is largely based on the references [IPO],
[KO 9], [KO 12], [KO 19-2] and [NR].
In addition, the following changes are made:
Section 2.2 describes the criteria to exclude particular substances from the QRA
calculations. Article 9, Paragraph 6 of Council Directive 96/82/EC indicates that particular
substances in a state incapable of creating a major-accident hazard can be excluded from the
Safety Report [EU96]. The criteria to be used are given in the Commission decision on
harmonized criteria for dispensations according to Article 9 of Council Directive 96/82/EC
of 9 December 1996 on the control of major-accident hazards involving dangerous
substances [EU98]. Consequently, it has been decided to exclude these substances from the
QRA calculations using the same criteria.
Section 2.3.2.1 records the rule on whether mixtures and preparations of toxic substances
need to be considered. This rule has been changed. Previously, dangerous substances in
concentrations less than 5% did not need to be considered. This rule is now replaced by the
limits of the corresponding EU directives [EU88].
To facilitate QRA calculations in case many different substances are stored at different
times, the use of sample substances has been added in Section 2.3.2.1.
The factor O
3
accounts for the process conditions and is a measure of the amount of
substance in the gas phase after the release (see Section 2.3.2.2.3). In the calculation of O
3
,
an amount is used to account for the extra evaporation due to the heat flux from the
environment to the liquid pool formed. The use of the amount deviates from the
calculations used previously [P 172, IPO]. In [P 172, IPO], the amount is only added if
the process temperature is lower than the ambient temperature. This condition is omitted
here for two reasons:
the addition of an amount is meant to account for the extra evaporation caused by the
heat flux of the environment to the liquid pool. Therefore it is more reasonable to have
the value of not dependent on the process temperature, but only on the difference
between the atmospheric boiling point and the (fixed) ambient temperature.
In practice, this condition is likely not to be tested. The saturation pressure at process
temperature for most substances will be higher than 3 bar if the process temperature is
equal to or higher than 25 C and the atmospheric boiling point is lower than 25 C. A
saturation pressure at a process temperature higher than 3 bar already results in the
maximum value of X = 10.
For substances in the liquid phase, a factor X is used to calculate the factor O
3
(see Section
2.3.2.2.3). The use of an interpolation for the factor X between 1 and 10 has not
previously been clearly described for dangerous substances in non-dangerous solvents. The
interpolation is introduced here to be more in line with pure substances.
The calculation of the limit value, G, has been copied from [SZW97] (see Section 2.3.2.3).
One modification is made: [SZW97] assigns a limit value of 1 kg to substances which are
extremely toxic. Therefore a number of carcinogenic substances have a limit value of 1 kg,
although acute effects are not known. Since the QRA is directed to short-term lethal effects,
the category of extremely toxic substances is no longer included.
The selection method for large inter-unit pipelines represents a new procedure not
previously described. The two-phase outflow is calculated using PHAST V5.2 [DNV98].
The outflow is calculated for a pipeline connected to a large spherical vessel at a height of
one metre. The mass of the pressurized liquefied gas in the vessel is equal to 500 ton, the
filling grade is 0.9 and temperature T = 282 K. The length of the pipeline emptied in 600
seconds is determined iteratively. First, a pipeline length is postulated. Next, the mass
released in 600 seconds following a rupture of the pipeline is calculated. Finally, the volume
corresponding to the mass released and the pipeline length corresponding with this volume
is calculated. Using the new pipeline length, the procedure is repeated until convergence
occurs.
Appendix 2.A describes to procedure to assess the obligation to make a Safety Report. The
procedure is taken from [EU96]. The note describing the presence of a substance in step 1
is not stated in [EU96] but taken from [KO 12].
LOCs 3.1
3. LOSS OF CONTAINMENT EVENTS
3.1 Introduction
This chapter describes the Loss of Containment events (LOCs) that need to be included in the
QRA for establishments. The complete set of LOCs consists of generic LOCs, external-impact
LOCs, loading and unloading LOCs and specific LOCs.
Generic LOCs
The generic LOCs cover all failure causes not considered explicitly, like corrosion, construction
errors, welding failures and blocking of tank vents.
External-impact LOCs
LOCs for external impact are considered explicitly for transport units. The external-impact LOCs
applying to stationary installations and pipelines are assumed to be either already included in the
generic LOCs or should be included by adding an extra failure frequency.
Loading and unloading LOCs
Loading and unloading LOCs cover the transhipment of material from transport units to
stationary installations and vice versa.
Specific LOCs
Specific LOCs cover the LOCs specific to the process conditions, process design, materials and
plant layout. Examples are runaway reactions and domino effects.
Only LOCs that contribute to the individual and/or societal risk should be included in the QRA.
This means that LOCs of an installation should be included only if two conditions are fulfilled:
i.e. (1) frequency of occurrence is equal to or greater than 10
-8
per year and (2) lethal damage (1%
probability) occurs outside the establishments boundary or the transport route.
The LOCs for establishments are described in Sections 3.2.1 - 3.2.9.
3.2 Loss of Containment events at establishments
Loss of Containment events (LOCs) are defined for various systems in an establishment. The
systems and their LOCs are described in more detail in the sections as indicated in Table 3.1.
LOCs 3.2
Table 3.1 LOCs for systems in an establishment
System Section
Stationary t anks and vessels, pressurised 3.2.1
Stationary t anks and vessels,
at mospheric
3.2.2
Gas cylinders 3.2.1
Pipes 3.2.3
Pumps 3.2.4
Heat exchangers 3.2.5
Pressure relief devices 3.2.6
Warehouses 3.2.7
Storage of explosives 3.2.8
Road t ankers 3.2.9
Tank wagons 3.2.9
Ships 3.2.9
3.2.1 Stationary pressurised tanks and vessels
Of the various types of pressurised stationary tanks and vessels, pressure, process and reactor
vessels can be distinguished. These are described below.
Pressure vessel
A pressure vessel is a storage vessel in which the pressure is (substantially) more than 1 bar
absolute.
Process vessel
In a process vessel a change in the physical properties of the substance occurs, e.g. temperature
or phase. Examples of process vessels are distillation columns, condensers and filters. Vessels
where only the level of liquid changes can be considered as pressure vessels.
Reactor vessel
In reactor vessels a chemical change of the substances occurs. Examples of reactor vessels are
batch and continuous reactors. A vessel where a strong exothermic mixing of substances occurs
should also be considered as a reactor vessel.
The LOCs for pressure, process and reactor vessels are given in Table 3.2, the failure frequencies
of these LOCs for stationary vessels in
Table 3.3.
Table 3.2 LOCs for stationary vessels
LOC for st ationary vessels
G.1 Instant aneous release of the complet e inventory
G.2 Continuous release of t he complet e inventory in 10 min at a constant rate of release
G.3 Continuous release from a hole with an effective diameter of 10 mm
LOCs 3.3
Table 3.3 Frequencies of LOCs for stationary vessels
Installat ion (part ) G.1
Instantaneous
G.2
Continuous,
10 min
G.3
Continuous,
10 mm
pressure vessel 5 10
-7
y
-1
5 10
-7
y
-1
1 10
-5
y
-1
process vessel 5 10
-6
y
-1
5 10
-6
y
-1
1 10
-4
y
-1
react or vessel 5 10
-6
y
-1
5 10
-6
y
-1
1 10
-4
y
-1
Notes:
1. A vessel or tank consists of the vessel (tank) wall and the welded stumps, mounting plates
and instrumentation pipes. The LOCs cover the failure of the tanks and vessels and the
associated instrumentation pipework. The failure of pipes connected to the vessels and
tanks should be considered separately (see Section 3.2.3).
2. The failure frequencies given here are default failure frequencies based on the situation that
corrosion, fatigue due to vibrations, operating errors and external impacts are excluded. A
deviation of the default failure frequencies is possible in specific cases.
A lower failure frequency can be used if a tank or vessel has special provisions
additional to the standard provisions, e.g. according to the design code, which have an
indisputable failure-reducing effect. However, the frequency at which the complete
inventory is released (i.e. the sum of the frequencies of the LOCs, G.1 and G.2) should
never be less than 1 10
-7
per year.
A higher frequency should be used if standard provisions are missing or under
uncommon circumstances. If external impact or operating errors cannot be excluded, an
extra failure frequency of 5 10
-6
per year should be added to LOC G.1,
Instantaneous and an extra failure frequency of 5 10
-6
per year should be added to
LOC G.2, Continuous, 10 min.
3. Vessels and tanks can be (partly) in-ground, or situated inside or outside a building. The
LOCs and their frequencies are not dependent on the situation. The modelling of a release
inside a building is described in Chapter 4.
4. Storage tanks can be used for the storage of different substances at different times. If large
numbers of different substances are transhipped from an establishment, it is useful to
classify the substances and use sample substances for each category in the QRA. A
classification method is described in [VVoW95]. It should be noted that if a specific
substance makes up an important part of the total amount transhipped, the substance itself
will have to be used in the calculation.
LOCs 3.4
5. Storage tanks may have a pressure just above 1 bar absolute. These tanks have to be
considered as atmospheric storage tanks. Examples are cryogenic tanks and atmospheric
storage tanks with nitrogen blanketing.
6. The potential consequences of simultaneous failure of more than one tank should be
considered. For instance, if several tanks are located close to each other, a BLEVE of one
tank may lead to the failure of several other tanks. If several tanks are located in one bund,
the capacity of the bund should be sufficient to contain the liquid of all tanks, otherwise
simultaneous failure of more than one tank may lead to a spill outside the bund.
7. Failure frequencies of process and reactor vessels are higher by a factor of 10 than the
failure frequencies of pressure vessels. This factor covers the hazards imposed by the
chemical process, like runaway reactions unidentified in the analysis of the process.
However, the process is assumed to be analysed using methods like HAZOP, what/if and
checklist analyses and appropriate measures are taken to prevent the hazards identified. A
more complete description of analysis methods is given in the Red Book [CPR12E].
8. Catastrophic failure of a gas cylinder does not generally lead to lethal effects outside the
establishment. However, the possibility of domino effects should be considered, e.g.
following catastrophic failure of a gas cylinder with acetylene. The frequency of
catastrophic failure of a gas cylinder (instantaneous release) is 1 10
-6
per year.
3.2.2 Stationary atmospheric tanks and vessels
The various types of stationary tanks and vessels can be distinguished as given below:
Single-containment atmospheric tank
A single-containment atmospheric tank consists of a primary container for the liquid. An outer
shell is either present, or not, but when present, primarily intended for the retention and
protection of insulation. It is not designed to contain liquid in the event of the primary
containers failure.
Atmospheric tank with a protective outer shell
An atmospheric tank with a protective outer shell consists of a primary container for the liquid
and a protective outer shell. The outer shell is designed to contain the liquid in the event of failure
of the primary container but is not designed to contain any vapour. The outer shell is not
designed to withstand all possible loads, e.g. explosion (static pressure load of 0.3 bar during 300
ms), penetrating fragments and cold (thermal) load.
Double-containment atmospheric tank
A double-containment atmospheric tank consists of a primary container for the liquid and a
secondary container. The secondary container is designed to contain the liquid in the event of
failure of the primary container and to withstand all possible loads, like explosion (static pressure
load of 0.3 bar during 300 ms), penetrating fragments and cold (thermal) load. The secondary
container is not designed to hold any kind of vapour.
Full-containment atmospheric tank
LOCs 3.5
A full-containment atmospheric tank consists of a primary container for the liquid and a
secondary container. The secondary container is designed to contain both the liquid and vapour in
the event of failure of the primary container, and to withstand all possible loads, like explosion
(static pressure load of 0.3 bar during 300 ms), penetrating fragments and cold. The outer roof is
supported by the secondary containment and designed to withstand loads e.g. explosion.
Membrane tank
A membrane tank consists of a primary and a secondary container. The primary container is
formed by a non-self-supporting membrane that holds the liquid and its vapour under normal
operating conditions. The secondary container is concrete and supports the primary container.
The secondary container has the capacity to contain all the liquid and to realise controlled venting
of the vapour if the inner tank fails. The outer roof forms an integral part of the secondary
containment.
In-ground atmospheric tank
An in-ground atmospheric tank is a storage tank in which the liquid level is at or below ground
level.
Mounded atmospheric tank
A mounded atmospheric tank is a storage tank that is completely covered with a layer of soil and
in which the liquid level is above ground level.
The LOCs for atmospheric tanks are given in Table 3.4 and the frequencies of these LOCs in
Table 3.5.
Table 3.4 LOCs for atmospheric tanks
LOCs for atmospheric t anks
a direct ly to the at mosphere
b from the primary container int o t he unimpaired secondary container or out er shell
G.2 Continuous release of t he complet e invent ory in 10 min at a const ant rat e of release
Full-containment atmospheric tank
LOCs 3.6
Table 3.5 Frequencies of LOCs for atmospheric tanks
Installation (part) G.1a
Instantan.
release to
atmosphere
G.1b
Instantan.
release to
secondary
container
G.2a
Continuous
10 min
release to
atmosphere
G.2b
Continuous
10 min
release to
secondary
container
G.3a
Continuous
10 mm
release to
atmosphere
G.3b
Continuous
10 mm
release to
secondary
container
single-
containment tank 5 10
-6
y
-1
5 10
-6
y
-1
1 10
-4
y
-1
tank with a
protective outer
shell 5 10
-7
y
-1
5 10
-7
y
-1
5 10
-7
y
-1
5 10
-7
y
-1
1 10
-4
y
-1
double
containment tank 1.25 10
-8
y
-1
5 10
-8
y
-1
1.25 10
-8
y
-
1
5 10
-8
y
-1
1 10
-4
y
-1
full containment
tank 1 10
-8
y
-1
membrane tank see note 7
in-ground tank 1 10
-8
y
-1
mounded tank 1 10
-8
y
-1
Notes:
1. A vessel or tank consists of the vessel (tank) wall and the welded stumps, mounting plates
and instrumentation pipes. The LOCs cover the failure of the tanks and vessels, and the
associated instrumentation pipework. The failure of pipes connected to the vessels and
tanks should be considered separately (see Section 3.2.3).
2. Tanks can be situated inside or outside a building. The LOCs are not dependent on the
situation. Modelling a release inside a building is described in Chapter 4.
3. Storage tanks can be used for storing different substances at different times. If large
numbers of different substances are transhipped from an establishment, it is useful to
classify the substances and use sample substances for each category in the QRA. A
classification method is described in [VVoW95]. It should be noted that if a specific
substance constitutes an important part of the total amount transhipped, the substance
itself will have to be used in the calculation.
4. A cryogenic tank is an atmospheric tank with a storage temperature below ambient
temperature. The LOCs for a cryogenic tank are the LOCs of the corresponding type of
atmospheric storage tank.
LOCs 3.7
5. Atmospheric storage tanks may have a pressure just above 1 bar absolute. These tanks
should be considered as atmospheric storage tanks. Examples are cryogenic tanks and
atmospheric storage tanks with nitrogen blanketing.
6. The potential consequences of simultaneous failure of more than one tank should be
considered. For instance, if several tanks are located in one bund, the capacity of the bund
should be sufficient to contain the liquid of all tanks, otherwise simultaneous failure of
more than one tank may lead to a spill outside the bund.
7. The failure frequency of a membrane tank, determined by the strength of the secondary
container, should be estimated case by case using the data on the other types of
atmospheric tanks.
8. The liquid level in an in-ground atmospheric tank is at or below ground level. The
surrounding soil should be considered as a secondary container; failure of the tank results in
flash and pool evaporation only.
3.2.3 Pipes
The LOCs for pipes cover all types of process pipes and inter-unit pipelines above ground of an
establishment. The LOCs for transport pipelines underground are given elsewhere. The LOCs for
pipes are given in Table 3.6 and LOC frequencies for pipes in Table 3.7.
Table 3.6 LOCs for pipes
LOCs for pipes
G.1 Full bore rupt ure
- out flow is from bot h sides of the full bore rupt ure
G.2 Leak
- out flow is from a leak with an effective diameter of 10% of t he
nominal diamet er, a maximum of 50 mm
Table 3.7 Frequencies of LOCs for pipes
Installation (part ) G.1
Full bore rupt ure
G.2
Leak
pipeline,
nominal diamet er < 75 mm 1 10
-6
m
-1
y
-1
5 10
-6
m
-1
y
-1
pipeline,
75 mm nominal diameter 150
mm
3 10
-7
m
-1
y
-1
2 10
-6
m
-1
y
-1
pipeline,
nominal diamet er > 150 mm 1 10
-7
m
-1
y
-1
5 10
-7
m
-1
y
-1
LOCs 3.8
Notes:
1. The figures given for the pipework failure rate are based on process pipework operating in
an environment where no excessive vibration, corrosion/erosion or thermal cyclic stresses
are expected. If there is a potential risk causing a significant leak, e.g. corrosion, a correction
factor of 3 - 10 should be applied, depending on the specific situation.
2. Pipes can be situated inside or outside a building. The LOCs are not dependent on the
situation. The modelling of a release inside a building is described in Chapter 4.
3. The location of the full bore rupture can be important to the outflow. If the location is
important, at least three full bore ruptures will have to be modelled:
upstream, i.e. situated directly at the vessel at the high pressure side with zero pipeline
length
middle, i.e. situated halfway along the pipeline
downstream, i.e. situated directly at the vessel at the low pressure side.
For short pipelines, less than 20 metres, the location of the full bore rupture is probably
not important; modelling one location for the full bore rupture, i.e. upstream, will suffice.
For the leak LOC the location of the leak is probably not that important to the outflow, so
that one location for the leak will suffice.
4. For long pipelines, failure locations have to be selected at regular distances to produce a
smooth risk contour. There should be enough locations to ensure that the risk contour does
not change substantially when the number of failure locations is increased. A reasonable
initial distance between two failure locations is 50 metres.
5. Failures of flanges are assumed to be included in the failure frequency of the pipeline; for
that reason, the minimum length of a pipe is set at 10 metres.
3.2.4 Pumps
The LOCs for pumps are given in Table 3.8 and the LOCs frequencies for pumps in Table 3.9.
Table 3.8 LOCs for pumps
LOCs for pumps
G.1 Cat ast rophic failure
- full bore rupt ure of the largest connect ing pipeline
G.2 Leak
- out flow is from a leak with an effective diameter of 10% of t he
nominal diamet er of the largest connect ing pipeline, wit h a maximum of 50
mm
LOCs 3.9
Table 3.9 Frequencies for pumps
Cat ast rophic
failure
G.2
Leak
pumps wit hout addit ional provisions 1 10
-4
y
-1
5 10
-4
y
-1
pumps wit h a wrought st eel containment 5 10
-5
y
-1
2.5 10
-4
y
-1
canned pumps 1 10
-5
y
-1
5 10
-5
y
-1
3.2.5 Heat exchangers
The LOCs for heat exchangers are given in Table 3.10 and LOCs frequencies for heat exchangers
in Table 3.11. Three different types of heat exchangers are:
heat exchangers where the dangerous substance is outside the pipes.
heat exchangers where the dangerous substance is inside the pipes, with the outer shell having
a design pressure higher than or equal to the maximum occurring pressure of the dangerous
substance inside the pipes
heat exchangers where the dangerous substance is inside the pipes, with the outer shell having
a design pressure less than the maximum occurring pressure of the dangerous substance inside
the pipes
Table 3.10 LOCs for heat exchangers
LOC for heat exchangers
G.2 Continuous release of t he complet e inventory in 10 min at a constant rate of release
G.4 Full bore rupture of t en pipes simultaneously
- out flow from bot h sides of t he full bore rupt ure
G.5 Full bore rupture of one pipe
G.6 Leak
- out flow from a leak wit h an effect ive diameter of 10% of the
nominal diamet er, with a maximum of 50 mm
LOCs 3.10
Table 3.11 Frequencies of LOCs for heat exchangers
Instantaneous
G.2
Continuous,
10 min
G.3
Continuous,
10 mm
heat exchanger,
dangerous substance out side pipes 5 10
-5
y
-1
5 10
-5
y
-1
1 10
-3
y
-1
Rupture,
10 pipes
G.5
Rupture,
1 pipe
G.6
Leak
heat exchanger,
dangerous substance inside pipes,
design pressure out er shell less than pressure of
dangerous substance
1 10
-5
y
-1
1 10
-3
y
-1
1 10
-2
y
-1
heat exchanger,
dangerous substance inside pipes,
design pressure out er shell more t han pressure of
dangerous substance
1 10
-6
y
-1
Note:
1. The releases are directly into the atmosphere. It is assumed that a contamination of the
cooling substance does not lead to external safety effects. If the heat exchanger is equipped
with safety devices, like a pressure relief valve, the operation of the safety equipment
should be considered in the determination of the outflow.
2. The outflow of the connecting pipelines should be considered as well.
3.2.6 Pressure relief devices
The opening of a pressure relief device results in an emission only if the device is in direct contact
with the substance and discharges directly to the atmosphere.
The LOC for pressure relief devices is given in Table 3.12 and LOCs frequency for pressure
devices in Table 3.13.
Table 3.12 LOCs for pressure relief devices
LOC for pressure relief devices
G.1 discharge of a pressure relief device with maximum discharge rate
LOCs 3.11
Table 3.13 Frequency of LOC for pressure relief devices
discharge
pressure relief device 2 10
-5
y
-1
3.2.7 LOCs for storage in warehouses
The LOCs for the storage of substances in warehouses concern both the handling of packaging
units and the possibility of fire in the warehouse.
These LOCs for the storage of substances in warehouses are given in Table 3.14 and the LOC
frequencies in Table 3.15. Both LOCs and calculation methods are described in more detail in
[CPR15].
Table 3.14 LOCs for the storage of substances in warehouses
LOCs for the storage of substances in warehouses
G.1 Handling solids: dispersion of a fraction of the packaging unit inventory as respirable
powder
G.2 Handling liquids: spill of the complete packaging unit inventory
S.1 Emission of unburned toxics and toxics produced in the fire
Table 3.15 Frequencies of LOCs for the storage of substances in warehouses
Installation (part )
G.1
Dispersion of
respirable powder
G.2
Liquid spill
S.1
Fire
storage of substances in warehouses
with prot ection levels 1 and 2
1 10
-5
per handling
of packaging unit
1 10
-5
per handling
of packaging unit
8.8 10
-4
y
-1
storage of substances in warehouses
with prot ection level 3
1 10
-5
per handling
of packaging unit
1 10
-5
per handling
of packaging unit
1.8 10
-4
y
-1
3.2.8 Storage of explosives
The LOCs for the storage of explosives are given in Table 3.16 and LOC frequencies for this
storage in Table 3.17.
LOCs 3.12
Table 3.16 LOCs for the storage of explosives
LOCs for the storage of substances in warehouses
G.1 mass detonation in a storage unit
G.2 fire in a storage unit
Table 3.17 Frequencies of LOCs for the storage of explosives
Installation (part )
G.1
Mass detonat ion
G.2
Fire
storage of explosives 1 10
-5
per year see not e 1
Notes:
1. If a detonation occurs in a storage unit, the LOC should be modelled as mass detonation in
a storage unit. If detonation is excluded, the LOC should be modelled as fire in a storage
unit.
2. The probability and effects of sympathetic detonation and spread of fire have to be
considered.
3. The calculation methods to assess the risks of explosive storage are described in the
following references:
Manual of NATO safety principles for the storage of military ammunition and
explosives (AC258), Allied Ammunition Storage and Transport Publication 1 (AASTP-
1), May 1992 [NATO92]. Relevant sections are Internal Safety (I-A-3 to I-A-24),
Air Blast ( II-5-15 to II-5-34) and Thermal Radiation (II-5-35 to II-5-40).
Committee for the Prevention of Disasters. Methods for the calculation of damage (the
Green Book). Voorburg: Ministry of Social Affairs and Employment, 1990
1
[CPR16].
Timmers, PGJ. Berekening van het in- en extern risico van explosievenopslag met
behulp van RISKANAL( Draft). Rijswijk: TNO, 1997 [Ti97].
The relevant sections of the NATO Manual AC258 are available for perusal at the Prins
Maurits Laboratory, TNO, Rijswijk. A copy of the relevant sections may be obtained
under certain conditions.
3.2.9 Transport units in an establishment
Transport units for loading and unloading activities may be present in an establishment. The
LOCs can be divided into: LOCs to cover the intrinsic failure of the transport unit, LOCs to
cover loading and unloading activities and LOCs to cover external impact due to accidents.
LOCs 3.13
3.2.9.1 Road tankers and tank wagons in an establishment
The LOCs for road tankers and tank wagons in an establishment are given in Table 3.18 and
frequencies of these LOCs in Table 3.19. A distinction is made between atmospheric and
pressurised tanks.
Table 3.18 LOCs for road tankers and tank wagons in an establishment
LOC for road t ankers and t ank wagons in an establishment
G.1 Instantaneous release of t he complet e inventory
G.2 Continuous release from a hole t he size of t he largest connect ion
- If the t ank is (part ly) filled with liquid, the release is modelled from the liquid phase
out of t he largest liquid connect ion.
L.1a Full bore rupture of t he loading/unloading hose
- The outflowis from both sides of the full bore rupt ure.
L.2a Leak of the loading/unloading hose
- The outflow is from a leak wit h an effect ive diamet er of 10% of the nominal
diamet er, wit h a maximum of 50 mm.
L.1b Full bore rupture of t he loading/unloading arm
- Out flow from both sides of the full bore rupture
L.2b Leak of the loading/unloading arm
- Outflow from a leak with an effective diameter of 10% of t he nominal
diamet er, wit h a maximum of 50 mm
E.1 Ext ernal impact
S.1 Fire under tank
- t o be modelled as an inst ant aneous release of t he complete inventory of t he tank
LOCs 3.14
Table 3.19 Frequencies of LOCs for road tankers and tank wagons in an establishment
G.1
Instantant.
G.2
Cont,
large conn.
L.1a
Full bore
hose
L.2a
Leak
hose
L.1b
Full bore
arm
L.2b
Leak
arm
E.1
Extern.
impact
S.1
Fire
tank,
pressurised 5 10
-7
y
-
1
5 10
-7
y
-
1
4 10
-6
h
-
1
4 10
-5
h
-
1
3 10
-8
h
-
1
3 10
-7
h
-
1
see note
1
see
note 2
tank,
atmospheric 1 10
-5
y
-
1
5 10
-7
y
-
1
4 10
-6
h
-
1
4 10
-5
h
-
1
3 10
-8
h
-
1
3 10
-7
h
-
1
see note
1
see
note 2
Notes:
1. The external impact LOCs for road tanker or tank wagon accidents in an establishment are
determined by the local situation. A calculation method is described elsewhere in this
report. In general, the LOCs for road tanker accidents do not have to be considered in an
establishment if measures have been taken to reduce road accidents, like speed limits.
2. Fire under a tank may lead to the instantaneous release of the complete inventory of the
tank. Various causes of failure may lead to a fire under a tank:
leakage of the connections under the tank followed by ignition. This event only occurs
for tanks loaded with flammable substances. The frequency is equal to 1 10
-6
per year
for pressurised tanks and 1 10
-5
per year for atmospheric tanks.
fire in the surroundings of the tank. The failure frequency is determined by the local
situation. Important aspects are the presence of tanks with flammable substances
nearby and failure during loading and unloading of flammable substances. A calculation
method to determine the failure frequency due to the presence of tanks with flammable
substances nearby is described elsewhere in this report.
3. LOCs are described here for transport units with large containers. Substances are also
transported in small packaging units like gas cylinders. LOCs can be considered for each
packaging unit separately. However, domino effects and the simultaneous failure of more
than one packaging unit due to external impact have to be considered.
3.2.9.2 Ships in an establishment
The LOCs for ships in an establishment cover loading and unloading activities, and external
impact. The LOCs for ships are given in Table 3.20 and LOC frequencies in Table 3.21.
LOCs 3.15
Table 3.20 LOCs for ships in an establishment
LOC for ships in an establishment
L.1 Full bore rupt ure of the loading/unloading arm
L.2 Leak of t he loading/unloading arm
- out flow from a leak wit h an effect ive diameter equal to 10% of the nominal
diamet er, wit h a maximum of 50 mm
E.1 Ext ernal impact , large spill
- gas t anker continuous release of 180 m
3
in 1800 s
- semi-gas t anker (refrigerated) continuous release of 126 m
3
in 1800 s
- single-walled liquid tanker continuous release of 75 m
3
in 1800 s
- double-walled liquid t anker continuous release of 75 m
3
in 1800 s
E.2 Ext ernal impact , small spill
- gas t anker continuous release of 90 m
3
in 1800 s
- semi-gas t anker (refrigerat ed) continuous release of 32 m
3
in 1800 s
- single-walled liquid tanker continuous release of 30 m
3
in 1800 s
- double-walled liquid t anker continuous release of 20 m
3
in 1800 s
Table 3.21 Frequencies of LOCs for ships in an establishment
Ship
L.1
Full bore
arm
L.2
Leak
arm
E.1
Ext ernal
large spill
E.2
Ext ernal
small spill
single-walled liquid t anker
6 10
-5
per t ranshipment
6 10
-4
per t ranshipment 0.1 f
0
0.2 f
0
double-walled liquid tanker 6 10
-5
per t ranshipment
6 10
-4
0
0.0015 f
0
gas tanker, semi-gas t anker 6 10
-5
per t ranshipment
6 10
-4
0
0.00012 f
0
* The base accident failure rate, f
0
, is equal to 6.7 10
-11
T t N, where T is the total number
of ships per year on the transport route or in the harbour, t the average duration of
loading/unloading per ship (in hours) and N, the number of transhipments per year (see note 1).
Notes:
1. The external impact LOCs for collision accidents of a ship are determined by the local
situation. If a ship is docked in a (small) harbour outside the transport routes, external-
impact LOCs do not have to be considered. However, if movement of ships near the ship
docked at the establishment is possible, collision LOCs do have to be considered. The
external-impact LOCs are calculated using the base accident rate, f
0
. The base accident
LOCs 3.16
failure rate, f
0
, is equal to 6.7 10
-11
T t N, where T is the total number of ships per
year on the transport route or in the harbour, t the average duration of loading/unloading per
ship (in hours) and N, the number of transhipments per year.
2. If a loading arm contains more than one pipe, a rupture of a loading arm corresponds to a
rupture of all pipes simultaneously.
LOCs 3.17
Appendix 3.A Commentary
3.A.1 LOCs to be included in the QRA
Only LOCs that contribute to the individual and/or societal risk should be included in the QRA
under the conditions that (1) the frequency of occurrence is equal to or greater than 10
-8
per year
and that (2) lethal damage (1% probability) occurs outside the establishments boundary or the
transport route. The conditions are taken from [IPO] with the exception of one change. In [IPO]
LOCs have to be included if the failure frequency is greater than 10
-8
per year and if lethal damage
(1%) in a residential area is possible. Consequently, the individual risk contours will depend on
the location of the residential areas. In an extreme case, if no residential areas are located near the
establishment, a risk contour is not necessarily calculated. The criterion therefore becomes: LOCs
should be included if the failure frequency is greater than 10
-8
per year and if lethal damage (1%
probability) outside the establishments boundary is possible. This criterion corresponds with
present-day practice. A threshold of 10
-8
per year as criterion for including LOCs is considered
reasonable since generic LOCs leading to the release of the complete inventory have failure
frequencies in the range 10
-5
to 10
-7
per year.
3.A.2 Failure data
3.A.2.1 General
The failure data given in this Section are copied from [IPO] unless otherwise indicated.
The failure data in [IPO] are largely based on the research done in the COVO study [COVO81].
Meanwhile, a number of review studies published [e.g. AM94, TNO98b, Ta98] show a tendency
for some systems towards higher failure frequencies than the ones reported here. However, an
update of the failure frequencies would require an extensive investigation into the original data
sources to determine the validity of the data and their applicability to current-day practice. Since
this investigation has not yet been accomplished, it was decided not to update the failure
frequencies in this document. Such an investigation, resulting in an update of the failure
frequencies, is anticipated in due time.
The failure frequencies given here do not take the quality of the management explicitly into
account. Various (international) projects have been initiated to assess the management system of
an establishment and to evaluate the quality of the management by applying management factors
to the failure frequencies. However, these projects have not resulted to date in a consistent
method to evaluate the management system; consequently, management factors are not
introduced in this document. The subject will be considered again with the update of the failure
frequencies.
3.A.2.2 Pressure vessels
The failure frequencies of pressure vessels determined in the COVO study are based on data
collected by Phillips and Warwick, Smith and Warwick, and Bush [COVO81, Ph69, Sm74,
Bu75]. The base failure rate of catastrophic rupture of a pressure vessel is set at 1 10
-6
per year
and is applicable to static, vibration free, pressure vessels operating under conditions of no
LOCs 3.18
corrosion (external or internal) and thermal cycling, i.e. typical storage pressure vessels. The base
failure rate of catastrophic rupture of process vessels and reactor vessels is assumed to be ten
times higher, i.e. 1 10
-5
per year. The failure rate of small leaks (a hole with an effective diameter
of 10 mm) is assumed to be ten times higher than the catastrophic failure rate.
An instantaneous release is not always found to lead to maximum effect distances. Hence,
catastrophic rupture is modelled partly as a continuous release of the complete inventory within
ten minutes. This LOC was previously defined as a release from a hole with an effective diameter
of 50 mm or, if the duration of this release exceeds ten minutes, a continuous release with a
duration of ten minutes [IPO]. For simplicity, the LOC is now defined as a release of the
complete inventory in ten minutes at a constant rate of release.
Previously, an LOC Serious leakage from a hole with an effective diameter of 50 mm with a
failure frequency of 1 10
-5
per year was defined in the COVO study and the IPO document of
1994 [COVO81, IPO]. This LOC was meant to cover the rupture of pipes connected to the
vessel. However, rupture of pipes is already covered by the LOCs of pipes (see Section 3.2.3).
To avoid double counting, the LOC Serious leakage from a hole with an effective diameter of 50
mm is left out (see also the corrections from RE-95 to the IPO document [IPO]).
The frequency of catastrophic rupture of a gas cylinder is set equal to the catastrophic failure
frequency of a pressure vessel. This frequency is in good agreement with the failure frequency of
explosion of a gas cylinder (i.e. 9 10
-7
per year) as reported in [AM94].
3.A.2.3 Atmospheric tanks
The failure frequencies of atmospheric tanks are based on expert judgement. The base failure rate
of catastrophic rupture of a single containment atmospheric tank is assumed to be ten times
higher than the base failure rate of catastrophic rupture of a storage pressure vessel, i.e. 1 10
-5
per year. The failure rate of small leaks (a hole with an effective diameter of 10 mm) is assumed to
be ten times higher than the catastrophic failure rate.
The effect of the various protection levels are weighed in the base failure rate of catastrophic
rupture.
The base failure rate of an atmospheric tank with a protective outer shell is assumed to be a
factor five less than the base failure rate of a single containment atmospheric tank, i.e.
2 10
-6
per year. It is assumed that in 50% of the catastrophic failures the protective outer
shell remains unimpaired and the release is into the unimpaired protective outer shell (failure
frequency 1 10
-6
per year). In the other 50% of the catastrophic failures, the protective
outer shell also fails and the release is directly to the environment (failure frequency 1 10
-6
per year).
The base failure rate of a double containment atmospheric tank is assumed to be a factor of 80
less than the base failure rate of a single containment atmospheric tank, i.e. 1.25 10
-7
per
year. It is assumed that in 80% of the catastrophic failures, the secondary container remains
unimpaired and the release is into the unimpaired secondary container (failure frequency
1 10
-7
per year). In the other 20% of the catastrophic failures, the secondary container also
fails and the release is directly to the environment (failure frequency 2.5 10
-8
per year).
The base failure rate of a full containment atmospheric tank (catastrophic failure of both inner
and outer containers) is assumed to be 1 10
-8
per year. Catastrophic failure of the inner tank
only does not lead to a release into the environment and is omitted here.
LOCs 3.19
The base failure rate of an in-ground atmospheric tank or a mounded atmospheric tank is
assumed to be equal to the base failure rate of a full-containment atmospheric tank, i.e 1 10
-8
per year. Failure of an in-ground atmospheric tank leads to pool evaporation only, whereas
failure of a mounded atmospheric tank leads to a release directly into the atmosphere.
The failure rate of small leaks of the primary container (a hole with an effective diameter equal to
10 mm) is assumed not to be influenced by the level of protection and is equal to 1 10
-4
per
year for all types of atmospheric tanks. However, small leaks of the primary container of a full
containment tank, a membrane tank, an in-ground tank and a mounded tank are assumed not to
result in release to the atmosphere and are thus omitted.
Similar to pressure vessels, catastrophic rupture is modelled partly as an instantaneous release
and partly as a continuous release within ten min. The subdivision does not apply if the base
failure rate of catastrophic rupture is equal to 1 10
-8
per year since the minimum frequency of
LOCs to be considered is equal to 1 10
-8
per year.
A tank roof collapse of cryogenic storage tanks leads to fragments of the roof falling into the
containment. The extra heat input results in an additional evaporation of the refrigerated
substance. This process is difficult to model and therefore usually not included in the QRA.
3.A.2.4 Pipes
The failure frequencies of pipes, copied from [IPO], are based on the COVO study [COVO81,
AEC75, SRS, Ph69, Sm74] and [Hu92].
The failure frequencies of catastrophic rupture of piping are given in the COVO study as:
diameter 50 mm failure frequency 1 10
-10
m
-1
h
-1
= 8.8 10
-7
m
-1
y
-1
50 < diameter 150 mm failure frequency 3 10
-11
m
-1
h
-1
= 2.6 10
-7
m
-1
y
-1
diameter > 150 mm failure frequency 1 10
-11
m
-1
h
-1
= 8.8 10
-8
m
-1
y
-1
The failure frequency of pipe rupture is given in [Hu92] as a function of the pipe diameter for
pipe diameters in the range of 50 - 250 mm:
10
log (failure rate per metre per year) = (0.0064 (pipe diameter in mm) + 5.56).
Note that the relation deviates from the line shown in the corresponding figure [IPO, Hu92].
Using this relation yields the following failure frequencies of pipe rupture:
diameter 50 mm failure frequency 1.3 10
-6
m
-1
y
-1
-7
m
-1
y
-1
-7
m
-1
y
-1
-8
m
-1
y
-1
The failure frequencies of catastrophic rupture of piping, copied from [IPO], therefore
corresponds to the failure frequencies used in the COVO study. The lower limit of the pipe
diameter classification is changed to 75 mm to agree with the failure frequencies given in [Hu92].
The COVO study yields the following figures for significant leakage (hole size between 5 - 15
mm depending on the pipeline diameter):
diameter 50 mm failure frequency leak = 10 failure frequency rupture
50 < diameter 150 mm failure frequency leak = 20 failure frequency rupture
diameter > 150 mm failure frequency leak = 30 failure frequency rupture
LOCs 3.20
The failure frequency of a leak is given in [Hu92] as a function of the hole diameter:
10
log (failure rate per metre per year) = (0.026 [hole diameter in mm] + 5.32)
If the hole diameter is set equal to 10% of the pipe diameter, the failure frequency of a leak
corresponds to:
diameter 50 mm failure frequency leak = 2.7 failure frequency rupture
The failure frequencies of a leak (effective diameter leak equal to 0.1 pipe diameter) are assumed
to be five times higher than the failure frequencies of catastrophic rupture [IPO]. This is in
reasonable agreement with the data in [Hu92].
3.A.2.5 Pumps
Pumps are not explicitly described in [IPO]. The COVO study gives 1 10
-4
per year as a
frequency of catastrophic failure of a pump [COVO81, SA75, SRS]. This failure frequency is
now used for pumps without additional provisions. Provisions are evaluated by expert
judgement.
The failure frequencies are averages, irrespective of pump type, type of drive, type of sealing,
number of revolutions, etc.
The catastrophic failure is modelled as a full bore rupture of the largest connecting pipeline.
Parallel to pipelines, a LOC leak is defined with a failure frequency equal to five times the
failure frequency of catastrophic failure.
3.A.2.6 Heat exchangers
Heat exchangers are not listed in [IPO]. The failure frequencies of heat exchangers listed here are
based on expert judgement only.
Similar to pressure vessels, catastrophic rupture of a heat exchanger with the dangerous substance
outside the pipes is modelled partly as an instantaneous release and partly as a continuous
release of the complete inventory within 10 minutes.
For heat exchangers with the dangerous substance inside the pipes, a rupture of 10 pipes is
assumed to always go simultaneously with failure of the outer shell and therefore results in a
direct release to the environment.
If the design pressure of the outer shell is higher than the maximum occurring pressure of the
dangerous substance in the pipes, rupture of one pipe and a leak is assumed not to lead to
emission outside the outer shell. Furthermore, a lower failure frequency is used to evaluate the
protective effect of the outer shell.
LOCs 3.21
3.A.2.7 Pressure relief devices
The LOC describes the opening of a pressure relief device and is based on expert judgement only
[IPO].
3.A.2.8 Storage
LOCs for the storage of substances in warehouses and their frequencies are copied from
[CPR15].
The LOCs for the storage of explosives are based on the present-day risk analysis method.
3.A.2.9 Transport units in an establishment
The LOCs for the intrinsic failure of a road tanker or tank wagon in an establishment are based on
expert judgement.
Catastrophic failure of a tank is modelled as an instantaneous release of the complete inventory.
Unlike stationary vessels, the LOC of catastrophic failure is not subdivided into a instantaneous
release and a continuous release within 10 minutes, since a continuous release is already modelled
as a release from a hole with the size of the largest connection.
The frequency of catastrophic failure of a pressurised tank is equal to 5 10
-7
per year. This
frequency represents a factor 2 less than the catastrophic failure frequency of a stationary
pressure vessel. It is therefore apparently assumed that the different conditions for road tankers
and tank wagons, e.g. the possibility of fatigue due to vibrations, are more than compensated by
measures taken in the construction of the tanks.
The frequency of catastrophic failure of an atmospheric tank is equal to 1 10
-5
per year. This
frequency is equal to the catastrophic failure frequency of a stationary single-containment
atmospheric tank. It is therefore apparently assumed that the different conditions for road
tankers and tank wagons, e.g. the possibility of fatigue due to vibrations, are compensated by
measures taken in the construction of the tanks.
The LOC Continuous release from a hole the size of the largest connection covers the failure of
connections at the tank and has a failure frequency of 5 10
-7
per year. The reasoning behind this
value is unknown.
If the tank is loaded with flammable materials, an additional LOC, S.1, is defined to cover a
catastrophic failure following leakage and ignition of the flammable substance. The overall failure
frequency of flammable material leakage, followed by ignition and in turn followed by an
instantaneous release of the complete inventory is set equal to 1 10
-6
per year for pressurised
tanks and 1 10
-5
per year for atmospheric tanks.
The frequency of catastrophic rupture of a loading arm or loading hose (road tankers and tank
wagons) is derived from the COVO study [COVO81, AEC75, We76, Ja71]. In this, the
frequency value for a lightly stressed hose is used; the frequency of a rupture of a heavily
LOCs 3.22
stressed hose is a factor 10 higher. The frequency of a leak is assumed to be 10 the frequency of
catastrophic rupture.
The failure frequencies of the LOCs for ships at an establishment are derived from [KO 22-5] and
[IPORBM]. The failure frequency of external impact is based on a number of accidents involving
bunker ships, leading to heavy damage. The derivation is described in [KO22-5]. The LOCs
associated with external impact are copied from [IPORBM].
The failure frequency during loading and unloading of ships is based on the number of spills in
the harbour of Rotterdam in the period 1976 - 1988, leading to a spill frequency of 6.7 10
-4
per
handled ship [KO 22-5]. The frequency of a leak is assumed to be 10 the failure frequency of a
rupture, leading to a frequency of rupture of 6 10
-5
per transhipment and a frequency of leakage
of 6 10
-4
per transhipment.
Model aspects 4.1
4. MODELLING SOURCE TERM AND DISPERSION
4.1 Introduction
After defining the loss of containment events, as described in Chapter 3, for a situation, the
source term and the dispersion in the environment will have to be calculated.
Models to calculate source term and dispersion are described extensively in the Yellow Book
[CPR14E] and in the risk analysis method for warehouses [CPR15]. Various types of models like
the following, are described:
outflow and spray release
pool evaporation
vapour cloud dispersion
jets and plumes
dense gas dispersion
passive dispersion
vapour cloud explosion
heat flux from fires
rupture of vessels
The models described in the Yellow Book are selected to combine good scientific performance
with ease of application in practice. For this reason these models are recommended for the QRA
calculations.
In the Netherlands, several integrated software packages are commonly used to assess the risks
associated with the storage, processing and transport of dangerous substances. In addition, more
complex models may be preferred for obtaining reliable results in specific situations.
Consequently, other models can be used for the QRA calculations besides those described in the
Yellow Book. However, the user should demonstrate adequate scientific performance in
applying the models. The scientific performance of the models should be demonstrated using the
results of validation exercises, model intercomparison studies and/or publications.
The models to be used in the QRA have already described elsewhere in detail [CPR14E]. For
each loss of containment event, an appropriate outflow model has to be selected. Next, various
parameters have to be set in the model calculation. This chapter describes the connection between
the loss of containment situations, as described in Chapter 3, and the models to calculate the
outflow and the dispersion in the environment. Subsequently, a number of recommended
parameter values are described.
It should be noted that plant-specific information should be used whenever possible. If plant-
specific information is not available, use can be made of the generic values given here.
Model aspects 4.2
4.2 Properties of substances
To calculate the source term and the dispersion of substances in the environment, the
(temperature-dependent) physical properties of the substances are needed. Information on the
physical properties can be found in a number of references and databases. Examples are:
The Yellow Book [CPR14E];
The DIPPR database [DIPPR];
Perry et al. [Pe84];
Reid et al. [Re88];
Yaws [Ya77].
4.3 Outflow models
Chapter 2 describes the loss of containment events (LOCs) to be used in the QRA. For each
LOC, the outflow of material can be calculated using the models in the Yellow Book. However,
there is still some freedom in the selection of the model to determine the outflow conditions.
Table 4.1 gives the correlation between the LOCs and the models described in the Yellow
Book.
Table 4.1 Correlation between LOCs and outflow models
Loss of cont ainment Inst allation To model as:
inst antaneous t anks and vessels
road tanker
t ank wagon
t otally rupt ured vessel
- gas: no air ent rainment during
expansion
- liquid: spreading pool
cont inuous release t anks and vessels
road tanker
t ank wagon
ships
hole in vessel wall (sharp orifice)
full bore rupt ure process pipes,
t ransport pipelines,
loading- /unloading
arm or hose
full bore rupt ured pipeline
leak process pipes,
t ransport pipelines
loading- /unloading
arm or hose
out flow t hrough small leak (sharp orifice)
emission of unburned t oxics
and t oxics produced in the
fire
Warehouses see [CPR15]
invent ory of a packaging unit
as respirable powder
spill of the complet e
invent ory of a packaging unit
pressure relief valve All hole in vessel wall (rounded orifice)
pool evaporat ion t anks and vessels pool evaporat ion
process scenarios t anks and vessels specific models
release inside building pipes, t anks, vessels see Section 4.6.3
Model aspects 4.3
Notes:
1. The Yellow Book does not describe a model for a totally ruptured vessel filled with
compressed gas and for a totally ruptured vessel filled with (non-boiling) liquid. In the case
of a totally ruptured vessel with compressed gas, it can be assumed that the initial cloud
expands isentropically to atmospheric pressure without entrainment of ambient air. In case
of a totally ruptured vessel with non-boiling liquid, it can be assumed that the liquid leads to
a spreading pool on the ground or water surface, growing from its original size, i.e. of the
ruptured vessel.
2. The continuous release from tanks, vessels and transport units is modelled as a hole in the
vessel wall with a sharp orifice. If the value of the discharge coefficient, C
d
, is not calculated
in the model, it should be set at C
d
= 0.62.
3. A leak in a pipe or loading/unloading arm/hose can be modelled as if a constant pressure
upstream is present. If the value of the discharge coefficient, C
d
, is not calculated in the
model, the value must be set to C
d
= 0.62.
4. If there is a full bore rupture in a pipeline, the value of the discharge coefficient, C
d
, must be
set to C
d
= 1.0 if the value is not calculated in the model.
5. Generic values are used for the pipeline characteristics if no additional information is
present, i.e.:
- no bends in the pipe
- a pipe wall roughness of 45 m
6. The discharge rate from a pressure relief valve is determined by the characteristics of the
pressure valve and the downstream piping. The discharge rate is set equal to the maximum
discharge rate.
7. The release duration is determined by the conditions of the installation and the type of
LOC. The release duration can be anywhere from instantaneous to several hours if no
counter-measures are taken. In the QRA calculation, the release duration is limited to a
maximum of 30 minutes; effects are calculated using only the mass released in the first 30
minutes following the start of the release to the environment.
8. If during normal operation the content of a vessel varies in time, this variation has to be
modelled. Discrete values are used for the vessel content and a LOC is subdivided into a
number of distinct situations. Each distinct situation is modelled with a specific value of
the vessel content. The percentage of occurrence of a specific vessel content is discounted
for in the frequency calculation.
9. The presence of pumps in pipelines and their volumetric flow have to be taken into
account in the calculation of the outflow. If no pump specifications are available, assuming
a release rate of 1.5 times the nominal pumping rate (increase due to loss of pressure head)
is suggested.
Model aspects 4.4
10. If the discharge is from the liquid section of the vessel, pure liquid is released. Flashing in
the hole is not modelled; flashing takes place outside the vessel.
11. The location of the release is determined by the specific situation. For instance, the location
of the pressure relief valve determines the location of its release. The outflow of a vessel or
tank can be modelled using a distribution of release locations at different heights. The failure
frequency of the vessel should then be divided over the various release locations. However,
care should be taken that all relevant processes are considered in the outflow calculation,
possibly requiring different models for the different release locations. As this may become a
time-consuming calculation, the following simplified and more conservative approach may
be applied:
One single release location is selected. The location of the release is assumed to be one
metre above ground level, but the total inventory of the tank or vessel is assumed to be
released.
If the vessel or tank is (partly) filled with liquid, the release should be modelled from the
liquid phase with a liquid head equal to half the maximum liquid head.
In process and reactor vessels, various substances can be present under different
conditions. For instance, in a distillation column a toxic substance can be present in the
gas phase, whereas a flammable solvent can be present in the liquid phase. In this case,
at least two release points have to be considered: (1) release of the toxic substance from
the gas phase and (2) release of the flammable substance from the liquid phase.
12. The direction of the release is determined by the specific situation. For instance, the
outflow of a pressure relief device is generally vertically oriented. If no specific information
is available, the direction of the outflow is set horizontal, parallel to the wind direction. An
exception to this rule is an underground pipeline, of which the direction of release is
vertical.
13. The outflow can be obstructed, for example, by the soil surface and objects in the direct
vicinity. The outflow is generally modelled as an unobstructed outflow. Obstruction of the
outflow should be modelled if the following two conditions are met:
1. the ratio L
o
/L
j
is less than 0.33, with L
o
the distance between the release point to the
obstacle and L
j
the length of the free jet
2. the probability P
i
that the ratio L
o
/L
j
is less than 0.33 should be larger than 0.5
considering all possible directions of the outflow.
The length L
j
of a free jet of gas can be estimated using the equation:
L
j
= 12 u
0
b
0
/ u
air
(4.1)
where:
u
0
velocity of the jet at the source (m s
-1
)
b
0
source radius (m)
u
air
average ambient wind velocity, set equal to 5 m s
-1
If the two conditions are met, an LOC with frequency, f, is divided in two separate LOCs:
- an obstructed outflow with frequency P
i
f
- an unobstructed outflow with frequency (1 P
i
) f
Model aspects 4.5
The obstructed outflow is modelled as a jet with the impulse reduced by a factor of 4. If
the outflow is obstructed by the ground, the release height should be set to 0.
4.4 Repression factors
Repression systems can be present to limit the release of substances into the environment. The
effect of repression systems can be taken into account in the QRA. A distinction is made
between blocking and other repression systems. In general, the effect of repression systems can
only be considered if the effectiveness of the system is demonstrated.
4.4.1 Blocking systems
Blocking systems can be present to limit the outflow once a loss of containment occurs. The
operation of blocking valves can be triggered by, for instance, a detection system for gas or the
detection of an excess flow. The blocking valves can be closed either automatically or through an
operator.
The effect of the blocking system is determined by various factors, such as the position of gas
detection monitors and the distribution thereof over the various wind directions, the detection
limit of the detection system, the system reaction time and the intervention time of an operator.
The effect of the blocking system should be determined in an event tree analysis; here, the failure
on demand of the system should be considered. The probability of failure on demand of the
system as a whole is about 0.01 per demand.
If the operation of the blocking system is not established through an analysis, default values can
be used, which here are to be considered as guidelines. It is assumed that an automatic detection
system is present, like a gas detection system of sufficient sensitivity with monitors covering all
wind directions. Three different types of blocking systems are distinguished, namely the
automatic blocking system, the remote-controlled blocking system and the hand-operated
blocking system.
An automatic blocking system is a system where the detection of the leakage and closure of
the blocking valves is fully automatic. There is no action of an operator required.
- The closing time of the blocking valves is two minutes.
- The failure upon demand for the blocking system is 0.001 per demand.
A remote-controlled blocking system is a system where the detection of the leakage is fully
automatic. The detection results in a signal in the control room. The operator validates the
signal and closes the blocking valves using a switch in the control room.
- The closing time of the blocking valves is ten minutes.
- The failure upon demand for the blocking system is 0.01 per demand.
A hand-operated blocking system is a system where the detection of the leakage is fully
automatic. The detection results in a signal in the control room. The operator validates the
signal, goes to the location of the blocking valves and closes the valves by hand.
- The closing time of the blocking valves is 30 minutes.
- The failure on demand for the blocking system is 0.01 per demand.
Model aspects 4.6
4.4.2 Other repression systems
Various repression systems can be installed to limit the effects following a loss of containment.
Examples are sprinkler installations to limit the spread of a fire, water shields to prevent the
dispersion of (water soluble) substances in the atmosphere and the use of foam to limit pool
evaporation.
The effect of a repression system may appear in the QRA calculations on the condition that the
effectiveness of the system is ascertained. The reaction time and effectiveness of the system must
be demonstrated with, for instance, data from the manufacturer and logbooks of tests and
exercises.
The effect of a repression system appears in the QRA using the following method:
1. Determine the time for the system to become effective, t
react.
2. Determine the effectiveness of the system.
3. Set the source term for the time period 0 to t
react
equal to the source term without the
repression system.
4. Correct the source term for the time period following t
react
with the effectiveness of the
repression system.
5. Take the failure upon demand of the repression system into account. The probability of
failure upon demand should be determined with tools like fault tree analysis. A default
value is 0.05 per demand.
4.5 Pool evaporation
Pool evaporation models are described in the Yellow Book. The following points have to be
considered:
The spread of the liquid pool is influenced by obstacles and provisions to discharge the spill.
Examples are bunds, inclined surfaces, outlet-pipes to storage tanks or discharge into the
sewerage. The presence of these provisions can be taken into account as a loss term for the
mass in the liquid pool or as a constraint to the dimensions of the liquid pool. However, the
consequences of the discharge of the liquid elsewhere has to be considered.
If a spill of liquid occurs in a bund, its characteristics have to be taken into account. If the
walls of the bund are sufficiently high, the bund prevents the spreading of the liquid pool and
the dimensions of the pool are restricted to those of the bund. An effective pool radius, R
pool
,
is then calculated from the bund area, A
bund
, using the equation:
R
pool
= (A
bund
/ ) (4.2)
Model aspects 4.7
Various physical properties of the soil are needed to calculate the spread of the liquid pool and
the pool evaporation. Default values are listed in Table 4.2 for the surface roughness and in
Table 4.3 for the thermodynamic properties.
Table 4.2 Characteristic average roughness length of the soil
Soil Average roughness length
(m)
flat sandy soil, concrete, stones, industrial site 0.005
normal sandy soil, gravel, railroad yard 0.010
rough sandy soil, farmland, grassland 0.020
very rough, sandy soil grown over and with potholes 0.025
Table 4.3 Heat conduction properties of various materials: thermal conductivity (
s
), density
(
s
), specific heat (C
p,s
) and thermal diffusivity (a
s
)
Material
s
(J s
-1
m
-1
K
-1
)
s
(kg m
-3
)
C
p,s
(J kg
-1
K
-1
)
a
s
(m
2
s
-1
)
isolation concrete 0.207 900 920 2.5 10
-7
light concrete 0.418 1800 920 2.5 10
-7
heavy concrete 1.3 2400 920 5.9 10
-7
clinkers 0.7 2000 836 4.2 10
-7
average subsoil 8 wt% moist 0.9 2500 836 4.3 10
-7
dry sandy subsoil 0.3 1600 799 2.0 10
-7
wet sand 8 wt% moist/clay 0.6 1940 937 3.3 10
-7
wood 0.2 550 2300 1.6 10
-7
gravel 2.5 2000 1140 11.0 10
-7
carbon steel 46.0 7840 460 128.0 10
-7
4.6 Vapour cloud dispersion
The models for the vapour cloud dispersion are described in the Yellow Book [CPR14E].
4.6.1 Coupling outflow and vapour cloud dispersion
The result of the outflow models as defined in Section 4.3 forms the input for the models to
calculate the vapour cloud dispersion. Since the outflow models result in a mass flow rate varying
in time, the time variation in the source term should be taken into account in the vapour cloud
dispersion calculation. Numerical integration methods are to be used as shown in the Yellow
Book [CPR14E].
The use of numerical integration methods may lead to elaborate and time-consuming calculations.
Therefore an approximation may be used in which the outflow is divided into a number of
discrete time segments. Next, the dispersion of the various segments has to be calculated using
dispersion models. This section gives some guidelines to divide a time-varying source term into
Model aspects 4.8
discrete time segments and to model the dispersion of the vapour cloud. However, it should be
noted that the modelling of the dispersion of time-varying releases is complicated and the rules
given here are to be used as guidelines only. For each time-varying source term consideration
should be given to what the best solution is.
A time-varying source term can be approximated through a number of discrete time segments
with constant outflow conditions by dividing the total mass released evenly over a number of
time segments. The outflow conditions in each time segment can be calculated using the following
rules:
Calculate the total mass released in the first 30 min following the LOC, M
rel
.
Decide on the number of time segments, N
seg
. A division into five time segments will suffice in
most calculations.
Divide the total mass released evenly over the time segments, i.e. the mass released in each
time segment, M
seg
, is equal to M
seg
= M
rel
/ N
seg
.
Calculate the release duration of the first time segment, D
rel.1
, as the time needed to release a
mass M
seg
.
Calculate the release rate in the first time segment, Q
rel.1
, as Q
rel.1
= M
seg
/ D
rel.1
.
Set the outflow conditions in the first time segment equal to the conditions corresponding with
the release rate Q
rel.1
.
Calculate the release duration, release rate and outflow conditions for the other time segments
using the same procedure.
Figure 4.1 shows an example where a time-varying release is approximated with five time
segments having equal mass.
It is preferred to use multiple time segments to approximate a time-varying release. However, a
time-varying source term can also be approximated with a single time segment having constant
outflow conditions, if modelling a single time segment can be shown to give results comparable to
a division in five time segments. The outflow conditions in a single time segment can be
determined using the following rules:
For flammables, the outflow conditions are equal to the conditions of the first time segment,
having approximated the time-varying release with five time segments. This means that the
release rate, Q
rel
, is equal to 20% of the total mass released, divided by the time needed to
release the first 20% of the total mass: Q
rel
= 0.2 M
rel
/ D
rel.1
. The release duration, D
rel
, is
equal to the total mass released, divided by the release rate: D
rel
= M
rel
/ Q
rel
. The outflow
conditions are equal to the conditions corresponding to the release rate Q
rel
.
For toxics, the outflow conditions are equal to the conditions of the second time segment,
having approximated the time-varying release with five time segments. This means that the
release rate, Q
rel
, is equal to 20% of the total mass released, divided by the time needed to
release the second 20% of the total mass: Q
rel
= 0.2 M
rel
/ D
rel.2
. The release duration, D
rel
, is
equal to the total mass released, divided by the release rate: D
rel
= M
rel
/ Q
rel
. The outflow
conditions are equal to the conditions corresponding with the release rate Q
rel
.
Model aspects 4.9
Figure 4.1 Approximation of a time-varying release with five time segments, each of which has
constant outflow conditions. The mass released is the same in each time segment.
Calculating the dispersion of various release time segments, originating, for example, from a time-
varying release or the release of a vapour cloud combined with pool evaporation, is complicated.
During the transport of the mass in the various time segments downwind, the clouds interfere due
to down-wind dispersion. In general, two different approaches can be followed:
Each cloud segment is treated as an independent puff release, not influenced by the leading or
trailing cloud segment. Dispersion of the puff occurs during the transport downwind, thus
increasing the passage time of the cloud and decreasing the concentration. This approximation
is probably suitable if a short-term, high initial release rate is followed by a long-term, small
release rate.
Each cloud segment is treated as an independent steady-state release. Dispersion of the puff
downwind does not occur during the transport downwind, and the passage time of the cloud is
equal to the release duration of the segment until the total release can be considered as
instantaneous. This approximation is probably suitable if the change in release rate between
two adjacent time segments is limited.
The toxic dose is calculated as the sum of the toxic doses of each cloud segment. In general, the
approximation of a cloud segment as an independent puff will overestimate the dispersion of the
actual cloud and thus underestimate the toxic dose received if the dose is more than proportional
to the concentration. The approximation of a cloud segment as an independent steady-state
release will underestimate the dispersion of the actual cloud and thus overestimate the toxic dose
received if the dose is more than proportional to the concentration.
The following guidelines are to be used to model multiple time segments in the release:
1. If two releases are present simultaneously, e.g. releases from both ends of a full pipe
rupture, the releases have to be added.
2. If two releases are sequential (e.g. as time segments of a time-varying release or as a cloud
from an evaporating pool following the vapour cloud that has drifted away) each cloud
segment is treated as an independent steady-state release neglecting the dispersion
downwind until the total release can be considered as instantaneous.
Model aspects 4.10
3. If rain-out and pool evaporation occurs, releases are present simultaneously and
sequentially. As long as the initial vapour cloud is present above an evaporating pool, the
mass released in the pool evaporation should be added to the mass in the vapour cloud. As
the vapour cloud has drifted away from the pool surface, the release from the evaporating
pool is treated as a new time segment.
4.6.2 Modelling the vapour cloud dispersion
The calculation of vapour cloud dispersion should at least take into account the processes:
jets and plumes
dense gas dispersion
passive dispersion
The wet and dry deposition processes can be modelled if the data are available, otherwise the
deposition processes do not have to be considered.
Chemical processes in the cloud need to be considered if the hazard of cloud release and
dispersion is strongly affected. An example is the release of HF, where the formation of dimers
and polymers, and the reaction with water vapour, changes the characteristics of the vapour
cloud.
The aerodynamical roughness length is a measure of the influence of the terrain on the cloud
dispersion. The use of the roughness length assumes that no large obstacles are present that will
affect the dispersion. The roughness length is defined by the terrain upwind. In general, a single
average value is used for the surroundings of the installation. Default values are given in Table 4.4.
Large obstacles change the dispersion of a cloud to a large extent, in which case the use of an
average roughness length is no longer applicable. The models described in the Yellow Book do
not account for the presence of obstacles. Some simple models exist for specific situations
[CPR14E]. However, the presence of large objects can only be addressed using complex
computer codes, like CFD calculations or wind tunnel experiments. If possible, the influence of
large obstacles should be determined quantitatively and be addressed in the QRA. If it is not
possible to quantify the influence of large obstacles, a qualitative description is required.
Dispersion models use an averaging time to calculate the maximum concentration and the plume
width. The values to be used for the averaging time t
av
are:
flammables t
av
= 20 sec
toxics t
av
= 600 sec
Model aspects 4.11
Table 4.4 Terrain classification in terms of aerodynamic roughness length, z
0
Class Short description of terrain z
0
(m)
1 open water, at least 5 km 0.0002
2 mud flats, snow; no vegetation, no obstacles 0.005
3 open flat terrain; grass, few isolated objects 0.03
4 low crops; occasional large obstacles, x/h > 20
(1)
0.10
5 high crops; scattered large obstacles, 15 < x/h < 20
(1)
0.25
6 parkland, bushes; numerous obstacles, x/h < 15
(1)
0.5
7 regular large obstacle coverage (suburb, forest) (1.0)
(2)
8 city centre with high- and low-rise buildings (3.0)
(2)
(1)
x is a typical upwind obstacle distance and h the height of the corresponding major
obstacles.
(2)
These values are rough indications. The use of an aerodynamic roughness length, z
0
,
does not account for the effects of large obstacles.
4.6.3 Release inside a building
Vessels, tanks and pipes can be situated inside a building. Following release, the dispersion of the
substances is affected by the building.
The following rules apply:
If the building is not expected to withstand the pressure load following the release of material,
the emission should be considered as entering directly into the atmosphere; the dispersion will
then be modelled as if no building is present.
If the building is expected to withstand the pressure load following the release of material, the
source term outside the building is derived from the source term inside the building and the
ventilation system. The location of the source is defined by the location of the outlet of the
mechanical ventilation and/or the location of the natural ventilation openings.
The release to the atmosphere is determined by the concentration inside the building as a function
of time and place. Models can be used to calculate the time- and space-dependent concentration
(see, for example, [Gi97]). However, often a more simple approach is used, assuming that the
concentration inside the building is uniform and instantaneously in equilibrium with the release
from the source. This simplifies the calculation of the release to the atmosphere considerably:
1. For a continuous release, the release to air, Q
out
(t), is set equal to the source term inside the
building, Q
in
(t). For an instantaneous release, the release to air, Q
out
, is given by the
equation:
Q
out
= M F / V; (4.3)
Model aspects 4.12
where:
Q
out
Source strength to the atmosphere (kg s
-1
)
M Mass released (kg)
V Volume of the room (m
3
)
F Ventilation rate (m
3
s
-1
)
The duration of the release, t
rel
, is equal to M / Q
out
.
2. The release is to be modelled as a continuous jet in cross wind [CPR14E]. The location of
the release is determined by the location of the ventilation opening.
3. If a release occurs within the recirculation zone downwind of the building or from the roof
of the building, the influence of the recirculation zone downwind of the building has to be
considered. The plume is entirely taken up in the lee and the concentration in the
recirculation zone can be expressed as:
C
rz
= Q
out
/ (K A u); (4.4)
C
rz
Concentration in the recirculation zone(kg m
-3
)
Q
out
Source strength to the atmosphere (kg s
-1
)
K Parameter to account for the building shape and orientation (-)
A Projection of the building in wind direction (m
2
)
u Wind speed at the height of the building (m s
-1
)
The parameter K depends on the shape and orientation of the building relative to the wind
direction and values of K range from K = 0.1 to K = 2. A default value of K = 1 can be
assumed in combination with an average projection of the building in the wind direction.
The length of the recirculation zone is assumed to be equal to three times the minimum of
either the width or the height of the building. To determine the concentration outside the
recirculation zone, a virtual source technique can be used, matching the concentration at the
end of the lee.
In case of a vertical release from a chimney at the top of the building, the influence of the
recirculation zone downwind of the building has to be considered. The uptake in the lee is
modelled in terms of a lowering of the release height and a change in the dispersion
coefficients. The model is described in [NNM98].
Note:
1. The modelling of a fire in a building equipped with smoke shutters is similar to the
modelling of warehouse fires (see Section 4.6.4). As long as the building is intact, the toxic
substances are assumed to be homogeneously mixed inside the building. The release is out
of the smoke shutters with atmospheric pressure and zero heat content. The effect of the
building wake on the release has to be considered. Once the building is no longer intact, it is
assumed that plume rise occurs and lethal effects are not expected anymore.
Model aspects 4.13
4.6.4 Fires and plume rise
In a fire, unburned toxics and toxic combustion products can be released to the environment. Due
to the high temperature of the cloud, the cloud will rise. The consequences of plume rise have to
be considered in the QRA.
In the case of fires within buildings, like CPR-15 storages, no plume rise is assumed to occur
during the first stage of the fire since the building is still intact and the building lowers the
temperature of the combustion products. After the first stage, plume rise occurs and
calculation is terminated as lethal effects are no longer expected [CPR15].
In the case of open fires, plume rise is assumed to occur immediately and no lethal effects are
expected.
4.7 Ignition
4.7.1 Direct ignition
The probability of direct ignition for stationary installations is given in Table 4.5 and for
transport units at an establishment in Table 4.6. For stationary installations, a division is made
into K1-liquid, low reactive gas and average/high reactive gas. The reactivity is listed in Table 4.7
for a number of substances.
Table 4.5 Probability of direct ignition for stationary installations
Source Substance
Continuous Instantaneous K1-liquid Gas,
low reactive
Gas,
average/high reactive
< 10 kg/s < 1000 kg 0.065 0.02 0.2
10 - 100 kg/s 1000 - 10,000 kg 0.065 0.04 0.5
> 100 kg/s > 10,000 kg 0.065 0.09 0.7
Table 4.6 Probability of direct ignition for transport units in an establishment
Source Probability of direct ignition
road tanker continuous 0.1
road tanker instantaneous 0.4
tank wagon continuous 0.1
tank wagon instantaneous 0.8
Model aspects 4.14
Table 4.7 Reactivity of a number of substances ([CPR14]). If little or no information is
available, substances are classified as high reactive. These substances are indicated
with an *.
Low Average High
(1)
1-chloro-2,3-epoxypropane 1,3-
dichloropropene
3-chloro-1-propene
ammonia
bromomethane
carbon monoxide
chloroethane
chloromethane
methane
tetraethyl lead
1-butene
1,2-diaminoethane
1,3-butadiene
acetaldehyde
acetonitrile
acrylonitril
butane
chloroethene
dimethylamine
ethane
ethene
ethylethanamine
formic acid
propane
propene
1-butanethiol*
acetylene
benzene*
carbon disulfide*
ethanethiol*
ethylene oxide
ethylformate*
formaldehyde*
hydrogensulfide*
methylacrylate*
methylformate*
methyloxirane*
naphtha, solvent*
tetrahydrothiophene*
vinylacetate*
Given an instantaneous release with direct ignition, a BLEVE and a fire ball may occur. The
probability of a BLEVE and fire ball, P
BLEVE
, is equal to:
stationary installations P
BLEVE
= 0.7
transport units in an establishment P
BLEVE
= 1.0
The mass in the BLEVE is set equal to the total inventory of the tank. The pressure at failure of
the vessel should be equal to 1.21 the opening pressure of the relief device or, if no relief device
is present, should be equal to the test pressure of the vessel.
Where no BLEVE and fire ball occur following an instantaneous release with direct ignition, a
vapour cloud expanded to atmospheric pressure and a liquid pool are formed. The direct ignition
of the vapour cloud is modelled as a flash fire and explosion, as described in Section 4.8. The
direct ignition of the liquid pool results in a pool fire. The mass in the vapour cloud depends on
the adiabatic flash fraction, , and is given by the relations in Table 4.8.
Table 4.8 Mass in vapour cloud following an instantaneous release with direct ignition
adiabatic flash fraction, Mass in vapour cloud
(fraction of the total inventory of the tank)
< 0.1 2
0.1 < 0.36 (0.8 0.028) / 0.26
0.36 1
Model aspects 4.15
4.7.2 Delayed ignition
To calculate the delayed ignition probability, various rules exist. Two different ways to apply
delayed ignition in the QRA calculation are described here, a calculation with actual ignition
sources (A) and a free field calculation (B).
A. Calculation with actual ignition sources
A QRA calculation can be done using the specific locations of the known ignition sources at the
establishment and outside the establishment. The distribution of ignition sources in the
environment should be known or can be anticipated. An overview of ignition sources and their
strengths is given in Appendix 4.A. It should be noted that if only a few (weak) ignition sources
are present, there is a probability that ignition of the cloud will not occur.
B. Free field calculation
A QRA calculation can be done using the specific locations of the known ignition sources at the
establishment. If the cloud is not ignited at the establishment, ignition is assumed to take place at
maximum cloud area, with cloud area defined as the surface area of the LFL-footprint of the
cloud. If an LFL-contour is not present outside the establishment, e.g. the spill of a flammable
liquid in a bund, and if ignition does not occur at the establishment, ignition is assumed not to
take place.
The Societal Risk calculation is to be done with calculation method A, the calculation with actual
ignition sources.
The Individual Risk calculation is to be done with either calculation method A or calculation
method B, to be decided by the competent authorities.
4.7.3 Substances both toxic and flammable
Substances that are both toxic and flammable should in principle be modelled using toxic
properties as long as the cloud is not ignited and flammable properties as soon as the cloud
ignites. However, this approach is currently too complicated for the models used. The LOC is
therefore split into two independent events, namely, a purely flammable event and a purely toxic
event.
Substances having a low reactivity (see Table 4.7) are to be modelled as a purely toxic event.
Examples are ammonia, carbon monoxide and tetraethyl lead.
Substances having an average or high reactivity (see Table 4.7) are to be modelled using two
independent events, namely, a purely flammable event and a purely toxic event. Examples are
acrolein, acrylonitril, allylalcohol and ethylene oxide. The probabilities of the flammable event and
the toxic event are determined by the probability of direct ignition, P
d.i
. If direct ignition occurs,
the event is modelled as a flammable event with probability P
d.i
. If direct ignition does not occur,
the event is modelled as a toxic event with probability (1 P
d.i.
). The toxic event is modelled as if
the substance is purely toxic, and the flammable event is modelled as if the substance is purely
flammable. To summarise, a LOC with frequency f is divided in two separate events:
Model aspects 4.16
- a purely toxic event with frequency (1 P
d.i.
) f.
Values of the probability of direct ignition, P
d.i.
, are given in Table 4.5 and Table 4.6.
Note:
1. Toxic effects after ignition of the flammable cloud are not considered. It is assumed that
after ignition plume rise occurs and toxic effects are no longer expected.
4.8 Effects of ignition of a vapour cloud
Following the ignition of an unconfined vapour cloud, one event occurs with characteristics from
both a flash fire and an explosion. This can be assumed to be modelled as two separate events,
namely, a pure flash fire and a pure explosion:
- a flash fire with no pressure effects, with a probability of 0.6;
- an explosion, with no flash fire effects, with a probability of 0.4.
The mass in the cloud is equal to the mass within the LFL-contour.
The side-on overpressure of the explosion can be calculated using the multi-energy method with
the highest value of the blast strength, 10 [CPR14E]. In the calculation, the flammable mass of
the cloud has to be partitioned into obstructed regions and non-obstructed regions, and the centre
of the explosion has to be determined by the location of the obstructed regions.
Assuming that a fraction f
obstr
of the total mass in the flammable cloud is in obstructed regions and
a fraction (1 f
obstr
) in non-obstructed regions, the distances to the peak side on overpressure
contours of 0.3 barg and 0.1 barg, R
0.3 barg
and R
0.1 barg
, respectively, can be calculated as:
R
0.3 barg
= 1.5 (f
obstr
E / P
a
)
1/3
(4.5)
R
0.1 barg
= 3 (f
obstr
E / P
a
)
1/3
(4.6)
with:
R
0.3 barg
distance to the peak side on overpressure contour of 0.3 barg(m)
R
0.1 barg
distance to the peak side on overpressure contour of 0.1 barg(m)
E combustion energy of the flammable mass within the LEL contour (J)
f
obstr
fraction of the total mass in the flammable cloud in obstructed regions (-)
P
a
ambient pressure (N m
-2
)
As default values, the fraction of the mass in obstructed regions, f
obstr
, can be set equal to 0.08,
and the centre of the explosion can be located at the centre of the cloud.
4.9 Rupture of vessels
The rupture of a pressurised vessel leads to the release of the internal energy, besides the release
of the vessel contents. The release of the internal energy can give rise to blast waves and high
velocity vessel fragments.
The release of the internal energy has to be considered in the determination of domino effects.
- a purely flammable event following direct ignition with frequency P
d.i.
f;
Model aspects 4.17
In general, the release of internal energy does not need to be considered in the determination of
effects outside the plant area.
4.10 Meteorological data
Meteorological data for the dispersion calculation can be expressed either in terms of the Monin-
Obukhov length L, or the Pasquill classes. The two types of classification can be related to each
other [CPR14E]. The use of the classification based on the Monin-Obukhov length L is
recommended, if appropriate statistical data are available. However, the classification used should
be consistent with the dispersion model and its dispersion coefficients.
In a QRA, at least six representative weather classes have to be used, covering the stability
conditions stable, neutral and unstable, and low and high wind speeds. In case a classification in
terms of Pasquill classes is used, at least the following six weather classes have to be covered:
Stability class Wind speed
(1)
B medium
D low
D medium
D high
E medium
F low
(1)
Low wind speed corresponds with 1 - 2 m s
-1
Medium wind speed corresponds to 3 - 5 m s
-1
High wind speed corresponds to 8 - 9 m s
-1
The number of wind directions should be at least eight. The statistics of the meteorological
situation should be deduced from a nearby, representative meteorological station. Some
information on the grouping of meteorological data in weather classes and the statistics of a
number of meteorological stations is included in Appendix 4.B.
The mixing height is normally not an important parameter in the calculation of lethal effects.
Values for the mixing height are given in the Yellow Book [CPR14E].
Default values for a number of meteorological parameters are given in Table 4.9. The values are
intended as yearly averaged values. If necessary, discrete values in time should be used to
distinguish between differences in day and night and between different periods in the year.
Table 4.9 Default values for a number of meteorological parameters
Parameter Default value
ambient air temperature 282 K
soil/bund temperature 282 K
water temperature 282 K
ambient pressure 101510 N/m
2
humidity 83 %
solar radiation flux 0.12 kW/m
2
Model aspects 4.18
The wind speed, temperature and pressure varies with the height above the surface [CPR14E].
The vertical variation of the wind speed should be addressed according to the calculation in the
Yellow Book. The minimum wind velocity is the wind velocity at a height of one metre.
Model aspects 4.19
Appendix 4.A Model to calculate the probability of delayed ignition
The probability of delayed ignition caused by an ignition source can be modelled as:
P(t) = P
present
(1 e
t
), (4.A.1)
where:
P(t) the probability of an ignition in the time interval 0 to t (-)
P
present
the probability that the source is present when the cloud passes (-)
the ignition effectiveness (s
-1
)
t time (s).
The ignition effectiveness, , can be calculated given the probability of ignition for a certain time
interval. Table 4.A.1 gives the probability of ignition for a time interval of one minute for a
number of sources. It should be noted, however, that the numbers are not well established and
should be used as a guideline.
Table 4.A.1 Probability of ignition for a time interval of one minute for a number of sources
Source Probability of ignition in one minute
Point Source
motor vehicle 0.4
flare 1.0
outdoor furnace 0.9
indoor furnace 0.45
outdoor boiler 0.45
indoor boiler 0.23
ship 0.5
ship transporting flammable materials 0.3
fishing vessel 0.2
pleasure craft 0.1
diesel train 0.4
electric train 0.8
Line source
transmission line 0.2 per 100 m
road Note 1
railway Note 1
Area source
chemical plant 0.9 per site
oil refinery 0.9 per site
heavy industry 0.7 per site
light industrial warehousing as for population
Population source
residential 0.01 per person
employment force 0.01 per person
Model aspects 4.20
Notes:
1. The ignition probability for a road or railway near the establishment or transport route
under consideration is determined by the average traffic density. The average traffic density,
d, is calculated as:
d = N E / v (4.A.2)
where:
N number of vehicles per hour (h
-1
)
E length of a road or railway section (km)
v average velocity of vehicle (km h
-1
).
If d 1, the value of d is the probability that the source is present when the cloud passes;
the probability of an ignition in the time interval 0 to t, P(t), equals:
P(t) = d (1 e
t
), (4.A.3)
where:
the ignition effectiveness of a single vehicle (s
-1
)
If d 1, d is the average number of sources present when the cloud passes; the probability
of an ignition in the time interval 0 to t, P(t), equals:
P(t) = (1 e
d t
), (4.A.4)
where:
the ignition effectiveness of a single vehicle (s
-1
)
2. The probability of an ignition for a grid cell in a residential area in the time interval 0 to t,
P(t), is given by:
P(t) = (1 e
n t
), (4.A.5)
where:
the ignition effectiveness of a single person (s
-1
)
n the average number of people present in the grid cell
3. Where the model uses a time-independent probability of ignition, the probability of ignition
is equal to the probability of ignition in one minute.
Model aspects 4.21
Appendix 4.B Meteorological data
Information on meteorological data is often available in terms of wind direction, wind speed and
stability classes. The information is usually expressed as fractional frequencies or numbers of
observation. To limit the calculation time for the QRA, it is useful to group the data in a limited
number of representative weather classes defined by wind speed and stability class.
It is recommended to use a classification of stability based on the Monin-Obukhov length L if
appropriate statistical data of a weather station nearby are available [NNM98, CPR14E].
However, these long-term statistical data may not yet be readily available, in which case Pasquill
stability classes may still be used. In this appendix an overview is given of the statistical data in
terms of Pasquill stability classes, based on routine meteorological observations as wind speed,
cloud cover and time of day.
In a QRA, at least six representative weather classes have to be used, covering the stability
conditions of stable, neutral and unstable, and low and high wind speeds. In terms of Pasquill
classes, at least the following six weather classes have to be covered (see Table 4.B.1).
Table 4.B.1 The six representative weather classes
Stability class Wind speed
(1)
B Medium
D Low
D Medium
D High
E Medium
F low
(1)
Low wind speed corresponds with 1 - 2 m s
-1
Medium wind speed corresponds to 3 - 5 m s
-1
High wind speed corresponds to 8 - 9 m s
-1
To group the observations in the six weather classes, the following rules apply:
1. Observations in the Pasquill stability classes A, A/B, B and B/C are allocated to stability
class B. The wind speed of the weather class is equal to the average wind speed of the
observations.
2. Observations in the Pasquill stability classes C, C/D and D, are allocated to stability class
D. Wind speeds below 2.5 m s
-1
(5 knots), between 2.5 m s
-1
and 6 m s
-1
(12 knots) and
above 6 m s
-1
are allocated to the wind speed categories low, medium and high, respectively.
The wind speed in each weather class is equal to the average wind speed of the
observations in the weather class.
3. Observations in the Pasquill stability classes E and F are allocated on the basis of the wind
speed. Wind speeds below 2.5 m s
-1
and above 2.5 m s
-1
are allocated to weather classes F
and E, respectively. The wind speed in each weather class is equal to the average wind
speed of the observations in the weather class.
Model aspects 4.22
The allocation is shown in Figure 4.B.1.
Figure 4.B.1 Allocation of observations into six weather classes.
Wind speed A B B/C C C/D D E F
< 2.5 m s
-1
D low F low
2.5 - 6 m s
-1
B medium D medium
E medium
> 6 m s
-1
D high
Data available can be separate for night-time and daytime, in which case, the period of the day
attributed to daytime should have the daytime and night-time statistics added correctly. One
should note that the population data are also divided in daytime and night-time. Care should be
taken to combine the population and weather data correctly.
Frequency distributions of a number of meteorological stations are listed in the following tables,
the data being compiled from [KNMI72]. The distributions with eight wind directions are derived
from the distributions with twelve wind directions using the conversions in Table 4.B.2.
The day period (i.e. day-time) refers to different time periods in different times of the year. For
the month of June, daytime corresponds to the hours of 6.00 - 21.00 MET and night-time to the
hours of 22.00 - 5.00 MET, whereas in December, daytime corresponds to the hours 10.00 -
16.00 MET and night-time to the hours of 17.00 - 9.00 MET. On average, daytime corresponds
to the time period 8:00 - 18:30 MET (fraction of 0.44) and night-time corresponds to 18:30 -
8:00 MET (fraction 0.56).
Direction 346-015 corresponds with wind coming from North.
Table 4.B.2 Conversion from the distribution with twelve wind directions to the distribution
with eight wind directions
Stability class Wind direction
B 3.0 m s
-1
B 4.0 m s
-1
D 1.5 m s
-1
D 1.5 m s
-1
D 5.0 m s
-1
D 4.0 m s
-1
D 9.0 m s
-1
D 8.0 m s
-1
E 5.0 m s
-1
F 4.0 m s
-1
F 1.5 m s
-1
F 1.5 m s
-1
fract ion (N-NO) = 0.5 fract ion (346-015) + fract ion (016-
045)
fract ion (NO-O) = fract ion (046-075) + 0.5 fract ion (076-
105)
......
...
Model aspects 4.23
Beek
Day B 3.0 m/s D 1.5 m/s D 5.0 m/s D 9.0 m/s E 5.0 m/s F 1.5 m/s Total
346-015 2.01 0.99 2.01 0.72 0.00 0.00 5.73
016-045 2.39 0.69 1.96 1.13 0.00 0.00 6.17
046-075 3.33 0.80 2.21 1.91 0.00 0.00 8.26
076-105 2.25 0.64 1.66 2.21 0.00 0.00 6.76
106-135 0.97 0.49 0.64 0.28 0.00 0.00 2.38
136-165 0.96 0.54 0.92 0.56 0.00 0.00 2.97
166-195 1.91 0.88 2.67 2.78 0.00 0.00 8.24
196-225 3.03 1.53 5.88 7.10 0.00 0.00 17.54
226-255 3.49 2.27 7.89 6.31 0.00 0.00 19.96
256-285 2.29 1.82 4.54 2.45 0.00 0.00 11.11
286-315 1.20 1.19 2.44 1.25 0.00 0.00 6.07
316-345 1.28 0.99 1.80 0.76 0.00 0.00 4.84
Total 25.11 12.83 34.61 27.46 0.00 0.00 100.00
Night B 3.0 m/s D 1.5 m/s D 5.0 m/s D 9.0 m/s E 5.0 m/s F 1.5 m/s Total
346-015 0.00 0.88 1.33 0.39 0.60 1.04 4.24
016-045 0.00 0.79 1.84 0.77 1.06 1.21 5.67
046-075 0.00 0.94 2.00 1.15 1.79 1.83 7.69
076-105 0.00 0.77 1.87 1.22 1.76 1.61 7.23
106-135 0.00 0.72 1.13 0.26 0.96 1.49 4.56
136-165 0.00 0.93 1.50 0.64 1.11 1.89 6.07
166-195 0.00 1.41 5.01 3.64 2.51 2.32 14.88
196-225 0.00 2.14 7.38 6.99 2.56 2.11 21.18
226-255 0.00 2.49 5.46 3.80 1.08 1.61 14.44
256-285 0.00 1.78 2.66 1.06 0.45 1.15 7.10
286-315 0.00 1.13 1.36 0.40 0.25 0.77 3.91
316-345 0.00 0.84 0.98 0.19 0.25 0.80 3.05
Total 0.00 14.80 32.51 20.48 14.38 17.83 100.00
Beek
Day N-NO NO-O O-ZO ZO-Z Z-ZW ZW-W W-NW NW-N Total
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 3.4 4.5 2.1 1.9 4.0 4.6 2.3 2.3 25.1
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 1.2 1.1 0.8 1.0 2.0 3.2 2.1 1.5 12.8
D 4.0 m/s 3.0 3.0 1.5 2.3 7.2 10.2 4.7 2.8 34.6
D 8.0 m/s 1.5 3.0 1.4 1.9 8.5 7.5 2.5 1.1 27.5
F 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 9.0 11.6 5.8 7.1 21.7 25.5 11.6 7.7 100.0
Night N-NO NO-O O-ZO ZO-Z Z-ZW ZW-W W-NW NW-N Total
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 1.2 1.3 1.1 1.6 2.8 3.4 2.0 1.3 14.8
D 4.0 m/s 2.5 2.9 2.1 4.0 9.9 6.8 2.7 1.6 32.5
D 8.0 m/s 1.0 1.8 0.9 2.5 8.8 4.3 0.9 0.4 20.5
F 1.5 m/s 1.7 2.6 2.3 3.1 3.3 2.2 1.3 1.3 17.8
F 4.0 m/s 1.4 2.7 1.8 2.4 3.8 1.3 0.5 0.5 14.4
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 7.8 11.3 8.2 13.5 28.6 18.0 7.5 5.2 100.0
Model aspects 4.24
Deelen
346-015 1.17 1.18 1.51 0.84 0.00 0.00 4.70
016-045 2.09 1.49 1.39 0.65 0.00 0.00 5.62
046-075 3.21 1.57 2.14 1.64 0.00 0.00 8.55
076-105 2.89 1.17 1.92 1.63 0.00 0.00 7.61
106-135 2.07 0.91 1.41 0.77 0.00 0.00 5.16
136-165 1.88 1.27 2.07 1.23 0.00 0.00 6.44
166-195 1.36 1.53 2.67 2.07 0.00 0.00 7.63
196-225 1.60 1.89 4.64 4.48 0.00 0.00 12.60
226-255 1.66 1.76 4.87 6.39 0.00 0.00 14.67
256-285 1.09 1.39 3.63 5.01 0.00 0.00 11.12
286-315 1.20 1.26 3.07 3.42 0.00 0.00 8.95
316-345 1.32 1.20 2.13 2.30 0.00 0.00 6.95
Total 21.54 16.61 31.44 30.43 0.00 0.00 100.00
346-015 0.00 1.37 0.71 0.19 0.30 2.35 4.91
016-045 0.00 1.50 1.10 0.47 0.64 2.76 6.47
046-075 0.00 1.84 2.68 1.45 2.18 3.35 11.50
076-105 0.00 1.38 2.27 1.01 1.73 3.49 9.88
106-135 0.00 1.66 1.51 0.41 1.23 4.20 9.01
136-165 0.00 1.54 1.88 1.04 0.62 2.39 7.47
166-195 0.00 1.72 2.28 1.75 0.45 1.53 7.73
196-225 0.00 2.12 3.76 3.49 0.87 2.13 12.36
226-255 0.00 1.97 3.74 4.26 0.80 1.69 12.45
256-285 0.00 1.60 2.55 2.26 0.61 1.38 8.40
286-315 0.00 1.37 1.32 0.99 0.29 1.20 5.16
316-345 0.00 1.33 0.92 0.42 0.21 1.78 4.66
Total 0.00 19.39 24.71 17.74 9.92 28.25 100.00
Deelen
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 2.7 4.7 3.5 2.6 2.3 2.2 1.7 1.9 21.5
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 2.1 2.2 1.5 2.0 2.7 2.5 2.0 1.8 16.6
D 4.0 m/s 2.1 3.1 2.4 3.4 6.0 6.7 4.9 2.9 31.4
D 8.0 m/s 1.1 2.5 1.6 2.3 5.5 8.9 5.9 2.7 30.4
F 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 8.0 12.4 9.0 10.3 16.4 20.2 14.5 9.3 100.0
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 2.2 2.5 2.3 2.4 3.0 2.8 2.2 2.0 19.4
D 4.0 m/s 1.5 3.8 2.6 3.0 4.9 5.0 2.6 1.3 24.7
D 8.0 m/s 0.6 2.0 0.9 1.9 4.4 5.4 2.1 0.5 17.7
F 1.5 m/s 3.9 5.1 5.9 3.2 2.9 2.4 1.9 3.0 28.2
F 4.0 m/s 0.8 3.0 2.1 0.8 1.1 1.1 0.6 0.4 9.9
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 8.9 16.4 14.0 11.3 16.2 16.7 9.4 7.1 100.0
Model aspects 4.25
Den Helder
346-015 0.52 0.25 1.10 4.81 0.00 0.00 6.68
016-045 0.71 0.28 1.02 4.34 0.00 0.00 6.36
046-075 1.80 0.37 1.69 5.01 0.00 0.00 8.87
076-105 1.43 0.36 1.93 3.38 0.00 0.00 7.10
106-135 0.96 0.40 1.43 1.37 0.00 0.00 4.15
136-165 0.73 0.52 1.36 0.49 0.00 0.00 3.10
166-195 1.21 0.71 2.59 3.26 0.00 0.00 7.77
196-225 0.73 0.46 1.98 11.30 0.00 0.00 14.47
226-255 1.17 0.38 2.32 9.79 0.00 0.00 13.67
256-285 1.29 0.44 1.91 7.28 0.00 0.00 10.92
286-315 1.20 0.37 1.32 5.13 0.00 0.00 8.02
316-345 1.09 0.36 1.43 6.03 0.00 0.00 8.91
Total 12.83 4.90 20.08 62.20 0.00 0.00 100.00
346-015 0.00 0.36 2.01 4.04 0.81 0.47 7.70
016-045 0.00 0.34 1.56 2.85 0.63 0.49 5.86
046-075 0.00 0.20 0.84 3.78 0.30 0.27 5.38
076-105 0.00 0.41 2.36 4.80 1.07 0.49 9.14
106-135 0.00 0.58 2.06 1.67 1.06 0.78 6.15
136-165 0.00 0.95 2.02 0.61 1.04 1.13 5.75
166-195 0.00 1.31 4.66 4.06 2.22 1.39 13.63
196-225 0.00 0.53 2.04 9.04 0.71 0.77 13.08
226-255 0.00 0.30 1.76 7.28 0.46 0.38 10.17
256-285 0.00 0.32 1.56 7.09 0.44 0.31 9.71
286-315 0.00 0.20 0.98 4.89 0.30 0.24 6.61
316-345 0.00 0.24 1.06 4.98 0.28 0.25 6.82
Total 0.00 5.75 22.89 55.08 9.31 6.97 100.00
Den Helder
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 1.0 2.5 1.7 1.3 1.3 1.8 1.8 1.4 12.8
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 0.4 0.5 0.6 0.9 0.8 0.6 0.6 0.5 4.9
D 4.0 m/s 1.6 2.7 2.4 2.7 3.3 3.3 2.3 2.0 20.1
D 8.0 m/s 6.7 6.7 3.1 2.1 2.9 13.4 8.8 8.4 52.2
F 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 9.7 12.4 7.7 7.0 8.4 19.1 13.5 12.2 90.0
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 0.5 0.4 0.8 1.6 1.2 0.5 0.4 0.4 5.7
D 4.0 m/s 2.6 2.0 3.2 4.3 4.4 2.5 1.8 2.1 22.9
D 8.0 m/s 4.9 6.2 4.1 2.6 11.1 10.8 8.4 7.0 55.1
F 1.5 m/s 0.7 0.5 1.0 1.8 1.5 0.5 0.4 0.5 7.0
F 4.0 m/s 1.0 0.8 1.6 2.2 1.8 0.7 0.5 0.7 9.3
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 9.7 10.0 10.7 12.6 19.9 15.0 11.5 10.7 100.0
Model aspects 4.26
Eelde
346-015 1.80 0.89 1.80 0.96 0.00 0.00 5.44
016-045 2.38 1.05 1.71 1.11 0.00 0.00 6.25
046-075 2.56 0.97 2.03 1.93 0.00 0.00 7.49
076-105 2.63 1.05 2.09 2.06 0.00 0.00 7.83
106-135 2.15 0.91 1.68 1.46 0.00 0.00 6.20
136-165 1.23 0.83 1.40 0.82 0.00 0.00 4.28
166-195 1.52 1.06 2.54 2.22 0.00 0.00 7.35
196-225 1.67 1.17 3.88 5.47 0.00 0.00 12.18
226-255 1.59 1.10 3.92 7.87 0.00 0.00 14.48
256-285 1.90 1.12 3.57 6.11 0.00 0.00 12.69
286-315 1.52 1.03 2.88 3.41 0.00 0.00 8.84
316-345 1.50 0.91 2.34 2.22 0.00 0.00 6.98
Total 22.43 12.09 29.85 35.63 0.00 0.00 100.00
346-015 0.00 0.91 0.74 0.29 0.33 1.35 3.62
016-045 0.00 1.19 0.99 0.32 0.66 2.25 5.41
046-075 0.00 1.15 2.00 1.43 1.34 2.84 8.76
076-105 0.00 1.22 2.22 1.51 1.54 2.65 9.15
106-135 0.00 1.41 1.77 0.98 0.90 2.22 7.27
136-165 0.00 1.24 1.45 0.74 0.54 1.67 5.63
166-195 0.00 1.49 2.68 2.04 0.94 2.01 9.16
196-225 0.00 1.76 4.59 4.52 1.64 2.55 15.07
226-255 0.00 1.52 3.96 5.15 1.57 2.34 14.54
256-285 0.00 1.71 2.80 2.68 1.12 2.56 10.87
286-315 0.00 1.40 1.53 1.19 0.42 1.84 6.38
316-345 0.00 0.90 1.14 0.64 0.28 1.20 4.15
Total 0.00 15.90 25.87 21.49 11.27 25.47 100.00
Eelde
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 3.3 3.9 3.5 2.0 2.4 2.5 2.5 2.4 22.4
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 1.5 1.5 1.4 1.4 1.7 1.7 1.6 1.4 12.1
D 4.0 m/s 2.6 3.1 2.7 2.7 5.1 5.7 4.7 3.2 29.8
D 8.0 m/s 1.6 3.0 2.5 1.9 6.6 10.9 6.5 2.7 35.6
F 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 9.0 11.4 10.1 7.9 15.8 20.8 15.2 9.7 100.0
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 1.6 1.8 2.0 2.0 2.5 2.4 2.3 1.4 15.9
D 4.0 m/s 1.4 3.1 2.9 2.8 5.9 5.4 2.9 1.5 25.9
D 8.0 m/s 0.5 2.2 1.7 1.8 5.5 6.5 2.5 0.8 21.5
F 1.5 m/s 2.9 4.2 3.5 2.7 3.6 3.6 3.1 1.9 25.5
F 4.0 m/s 0.8 2.1 1.7 1.0 2.1 2.1 1.0 0.4 11.3
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 7.2 13.3 11.8 10.2 19.6 20.0 11.8 6.0 100.0
Model aspects 4.27
Eindhoven
346-015 1.76 1.03 1.88 1.39 0.00 0.00 6.06
016-045 2.28 1.28 1.93 1.04 0.00 0.00 6.53
046-075 2.91 0.92 2.08 1.77 0.00 0.00 7.69
076-105 2.41 0.81 1.57 1.55 0.00 0.00 6.34
106-135 1.90 0.81 1.57 1.13 0.00 0.00 5.41
136-165 1.56 1.07 1.36 0.57 0.00 0.00 4.56
166-195 1.43 1.20 2.36 2.07 0.00 0.00 7.06
196-225 1.58 1.41 3.82 6.28 0.00 0.00 13.08
226-255 1.73 1.50 4.86 9.23 0.00 0.00 17.32
256-285 1.24 1.30 3.51 5.76 0.00 0.00 11.81
286-315 1.12 0.86 2.35 3.23 0.00 0.00 7.56
316-345 1.23 0.94 2.10 2.31 0.00 0.00 6.58
Total 21.15 13.14 29.39 36.32 0.00 0.00 100.00
346-015 0.00 0.83 1.00 0.42 0.60 1.84 4.69
016-045 0.00 1.40 1.44 0.60 0.95 2.73 7.11
046-075 0.00 1.14 2.00 1.03 1.53 2.90 8.61
076-105 0.00 0.80 1.47 1.04 1.17 1.83 6.31
106-135 0.00 1.27 1.60 0.80 1.00 2.38 7.05
136-165 0.00 1.54 1.69 0.56 0.81 2.46 7.05
166-195 0.00 1.80 2.56 1.75 0.88 2.47 9.45
196-225 0.00 1.89 4.05 5.10 1.33 2.41 14.77
226-255 0.00 1.76 4.41 6.31 1.22 1.78 15.49
256-285 0.00 1.48 2.54 2.82 0.82 1.68 9.33
286-315 0.00 1.08 1.39 1.04 0.49 1.45 5.45
316-345 0.00 0.87 1.15 0.56 0.39 1.71 4.69
Total 0.00 15.84 25.29 22.04 11.20 25.63 100.00
Eindhoven
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 3.2 4.1 3.1 2.3 2.3 2.4 1.7 2.1 21.1
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 1.8 1.3 1.2 1.7 2.0 2.2 1.5 1.5 13.1
D 4.0 m/s 2.9 2.9 2.4 2.5 5.0 6.6 4.1 3.0 29.4
D 8.0 m/s 1.7 2.5 1.9 1.6 7.3 12.1 6.1 3.0 36.3
F 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 9.6 10.9 8.6 8.1 16.6 23.2 13.5 9.6 100.0
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 1.8 1.5 1.7 2.4 2.8 2.5 1.8 1.3 15.8
D 4.0 m/s 1.9 2.7 2.3 3.0 5.3 5.7 2.7 1.6 25.3
D 8.0 m/s 0.8 1.6 1.3 1.4 6.0 7.7 2.5 0.8 22.0
F 1.5 m/s 3.6 3.8 3.3 3.7 3.6 2.6 2.3 2.6 25.6
F 4.0 m/s 1.3 2.1 1.6 1.2 1.8 1.6 0.9 0.7 11.2
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 9.5 11.8 10.2 11.8 19.5 20.2 10.1 7.0 100.0
Model aspects 4.28
Gilze-Rijen
346-015 2.06 1.24 2.11 0.98 0.00 0.00 6.39
016-045 2.90 1.35 2.37 1.51 0.00 0.00 8.14
046-075 2.67 0.94 2.07 2.30 0.00 0.00 7.98
076-105 1.53 0.66 1.33 1.72 0.00 0.00 5.24
106-135 1.46 0.68 1.31 1.06 0.00 0.00 4.51
136-165 1.20 0.81 1.44 0.70 0.00 0.00 4.14
166-195 1.18 0.97 2.50 2.51 0.00 0.00 7.16
196-225 1.74 1.45 4.70 5.71 0.00 0.00 13.60
226-255 2.01 1.67 5.14 7.20 0.00 0.00 16.01
256-285 1.99 1.63 4.02 5.10 0.00 0.00 12.74
286-315 1.55 1.41 3.14 2.24 0.00 0.00 8.34
316-345 1.30 1.05 2.22 1.17 0.00 0.00 5.74
Total 21.59 13.87 32.34 32.20 0.00 0.00 100.00
346-015 0.00 1.36 1.06 0.30 0.96 3.02 6.70
016-045 0.00 1.43 1.62 0.65 1.29 3.47 8.45
046-075 0.00 1.06 1.81 1.32 1.24 2.37 7.79
076-105 0.00 0.72 1.00 0.85 0.62 1.20 4.38
106-135 0.00 0.91 1.30 0.62 0.65 1.47 4.94
136-165 0.00 1.08 1.43 0.66 0.64 1.98 5.79
166-195 0.00 1.43 2.93 2.20 1.06 1.92 9.54
196-225 0.00 2.21 4.58 4.47 1.66 2.87 15.79
226-255 0.00 2.40 4.44 4.96 1.69 3.33 16.81
256-285 0.00 2.02 2.24 1.95 0.81 2.98 9.99
286-315 0.00 1.44 1.37 0.60 0.41 1.88 5.70
316-345 0.00 1.05 0.80 0.25 0.35 1.67 4.13
Total 0.00 17.10 24.56 18.81 11.37 28.16 100.00
Gilze-rijen
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 3.9 3.4 2.2 1.8 2.3 3.0 2.5 2.3 21.6
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 2.0 1.3 1.0 1.3 1.9 2.5 2.2 1.7 13.9
D 4.0 m/s 3.4 2.7 2.0 2.7 5.9 7.2 5.2 3.3 32.3
D 8.0 m/s 2.0 3.2 1.9 2.0 7.0 9.7 4.8 1.7 32.2
F 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 11.3 10.6 7.1 7.7 17.2 22.4 14.7 8.9 100.0
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 2.1 1.4 1.3 1.8 2.9 3.4 2.4 1.7 17.1
D 4.0 m/s 2.1 2.3 1.8 2.9 6.0 5.6 2.5 1.3 24.6
D 8.0 m/s 0.8 1.7 1.0 1.8 5.6 5.9 1.6 0.4 18.8
F 1.5 m/s 5.0 3.0 2.1 2.9 3.8 4.8 3.4 3.2 28.2
F 4.0 m/s 1.8 1.5 1.0 1.2 2.2 2.1 0.8 0.8 11.4
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 11.8 10.0 7.1 10.6 20.6 21.8 10.7 7.5 100.0
Model aspects 4.29
Hoek van Holland
346-015 2.36 0.67 2.75 5.01 0.00 0.00 10.79
016-045 1.18 0.49 1.77 2.33 0.00 0.00 5.77
046-075 1.25 0.70 1.71 1.61 0.00 0.00 5.26
076-105 2.86 0.99 2.24 1.77 0.00 0.00 7.85
106-135 1.35 0.60 1.38 1.14 0.00 0.00 4.47
136-165 1.60 0.79 1.81 1.56 0.00 0.00 5.77
166-195 1.00 0.70 2.46 3.77 0.00 0.00 7.92
196-225 0.62 0.47 1.97 6.31 0.00 0.00 9.37
226-255 1.25 0.48 2.42 11.38 0.00 0.00 15.53
256-285 2.01 0.65 2.51 6.12 0.00 0.00 11.29
286-315 1.63 0.69 1.82 3.91 0.00 0.00 8.05
316-345 1.69 0.64 1.85 3.77 0.00 0.00 7.94
Total 18.77 7.87 24.69 48.66 0.00 0.00 100.00
346-015 0.00 0.44 1.48 2.73 0.43 0.49 5.57
016-045 0.00 0.84 1.82 1.58 1.23 1.09 6.57
046-075 0.00 1.32 2.37 1.67 1.82 2.25 9.42
076-105 0.00 1.67 2.92 1.31 2.92 2.77 11.58
106-135 0.00 0.77 1.62 0.90 0.95 1.34 5.56
136-165 0.00 0.87 2.30 1.70 0.85 1.23 6.96
166-195 0.00 1.06 3.37 4.14 1.23 1.26 11.07
196-225 0.00 0.51 2.31 6.55 0.63 0.57 10.58
226-255 0.00 0.39 1.94 8.50 0.54 0.46 11.82
256-285 0.00 0.46 1.80 5.41 0.35 0.38 8.39
286-315 0.00 0.36 1.33 4.41 0.30 0.43 6.82
316-345 0.00 0.37 1.33 3.28 0.32 0.39 5.68
Total 0.00 9.07 24.56 42.18 11.54 12.65 100.00
Hoek van Holland
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 2.4 2.7 2.8 2.1 1.1 2.2 2.6 2.9 18.8
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 0.8 1.2 1.1 1.1 0.8 0.8 1.0 1.0 7.9
D 4.0 m/s 3.1 2.8 2.5 3.0 3.2 3.7 3.1 3.2 24.7
D 8.0 m/s 4.8 2.5 2.0 3.4 8.2 14.4 7.0 6.3 48.7
F 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 11.2 9.2 8.4 9.7 13.3 11.2 13.7 13.3 100.0
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 1.1 2.2 1.6 1.4 1.0 0.6 0.6 0.6 9.1
D 4.0 m/s 2.6 3.8 3.1 4.0 4.0 2.8 2.2 2.1 24.6
D 8.0 m/s 2.9 2.3 1.6 3.8 8.6 11.2 7.1 4.6 42.2
F 1.5 m/s 1.3 3.6 2.7 1.9 1.2 0.6 0.6 0.6 12.6
F 4.0 m/s 1.4 3.3 2.4 1.5 1.2 0.7 0.5 0.5 11.5
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 9.4 15.2 11.4 12.5 16.1 16.0 11.0 8.5 100.0
Model aspects 4.30
IJmuiden
346-015 0.87 0.48 1.80 3.94 0.00 0.00 7.09
016-045 0.71 0.41 1.57 1.84 0.00 0.00 4.53
046-075 1.15 0.37 1.59 3.10 0.00 0.00 6.21
076-105 2.11 0.54 2.74 4.39 0.00 0.00 9.77
106-135 1.25 0.65 1.62 1.60 0.00 0.00 5.11
136-165 0.86 0.67 1.96 2.04 0.00 0.00 5.51
166-195 0.58 0.58 1.99 2.99 0.00 0.00 6.13
196-225 0.91 0.52 2.31 8.06 0.00 0.00 11.80
226-255 1.65 0.57 2.65 9.95 0.00 0.00 14.82
256-285 1.50 0.55 2.08 7.11 0.00 0.00 11.24
286-315 1.24 0.51 1.83 5.29 0.00 0.00 8.86
316-345 1.41 0.49 2.16 4.85 0.00 0.00 8.91
Total 14.23 6.33 24.29 55.15 0.00 0.00 100.00
346-015 0.00 0.56 1.39 1.69 0.48 0.45 4.58
016-045 0.00 0.73 2.59 1.70 1.35 0.80 7.17
046-075 0.00 0.40 1.76 3.11 1.05 0.69 7.02
076-105 0.00 0.85 3.47 4.35 2.42 1.37 12.45
106-135 0.00 1.09 2.28 1.48 1.33 1.15 7.33
136-165 0.00 1.11 3.05 2.32 1.39 1.23 9.09
166-195 0.00 0.97 2.80 3.62 1.19 0.88 9.45
196-225 0.00 0.36 1.77 6.89 0.53 0.42 9.96
226-255 0.00 0.31 1.55 8.04 0.37 0.40 10.67
256-285 0.00 0.25 1.45 6.91 0.40 0.32 9.32
286-315 0.00 0.26 1.17 5.36 0.30 0.26 7.34
316-345 0.00 0.29 1.14 3.71 0.23 0.27 5.63
Total 0.00 7.18 24.40 49.17 11.03 8.23 100.00
IJmuiden
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 1.1 2.2 2.3 1.1 1.2 2.4 2.0 1.8 14.2
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 0.6 0.6 0.9 1.0 0.8 0.8 0.8 0.7 6.3
D 4.0 m/s 2.5 3.0 3.0 2.9 3.3 3.7 2.9 3.1 24.3
D 8.0 m/s 3.8 5.3 3.8 3.5 9.6 13.5 8.8 6.8 55.2
F 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 8.1 11.1 10.0 8.6 14.9 20.4 14.5 12.5 100.0
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 1.0 0.8 1.5 1.6 0.8 0.4 0.4 0.6 7.2
D 4.0 m/s 3.3 3.5 4.0 4.4 3.2 2.3 1.9 1.8 24.4
D 8.0 m/s 2.5 5.3 3.7 4.1 8.7 11.5 8.8 4.6 49.2
F 1.5 m/s 1.0 1.4 1.8 1.7 0.9 0.6 0.4 0.5 8.2
F 4.0 m/s 1.6 2.3 2.5 2.0 1.1 0.6 0.5 0.5 11.0
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 9.5 13.2 13.6 13.8 14.7 15.3 12.0 7.9 100.0
Model aspects 4.31
Leeuwarden
346-015 1.42 0.77 2.07 1.87 0.00 0.00 6.13
016-045 1.92 0.84 2.13 2.33 0.00 0.00 7.22
046-075 1.84 0.88 1.88 2.48 0.00 0.00 7.08
076-105 2.23 0.97 1.91 2.58 0.00 0.00 7.69
106-135 1.25 0.74 1.45 1.02 0.00 0.00 4.45
136-165 1.13 0.82 1.74 0.92 0.00 0.00 4.62
166-195 1.77 1.45 2.93 2.56 0.00 0.00 8.70
196-225 1.70 1.50 3.89 5.91 0.00 0.00 13.00
226-255 1.47 1.17 3.37 6.87 0.00 0.00 12.87
256-285 1.49 0.91 2.93 6.75 0.00 0.00 12.09
286-315 1.12 0.69 2.10 4.28 0.00 0.00 8.19
316-345 1.28 0.68 2.35 3.66 0.00 0.00 7.96
Total 18.63 11.42 28.75 41.21 0.00 0.00 100.00
346-015 0.00 0.85 1.17 0.76 0.47 1.23 4.48
016-045 0.00 1.02 1.28 0.72 0.71 1.63 5.36
046-075 0.00 1.15 1.74 1.43 1.27 2.33 7.92
076-105 0.00 1.17 2.19 2.08 1.77 2.48 9.68
106-135 0.00 0.87 1.73 1.05 0.99 1.38 6.03
136-165 0.00 1.08 2.20 1.07 0.82 1.21 6.37
166-195 0.00 1.85 3.52 2.40 1.73 2.74 12.23
196-225 0.00 2.00 4.05 4.84 1.72 3.10 15.71
226-255 0.00 1.40 2.82 3.78 1.03 2.09 11.11
256-285 0.00 1.10 2.10 3.26 0.84 1.55 8.85
286-315 0.00 0.87 1.61 2.40 0.70 1.03 6.60
316-345 0.00 0.90 1.46 1.79 0.44 1.06 5.65
Total 0.00 14.25 25.87 25.57 12.50 21.82 100.00
Leeuwarden
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 2.6 3.0 2.4 2.0 2.6 2.2 1.9 2.0 18.6
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 1.2 1.4 1.2 1.5 2.2 1.6 1.2 1.1 11.4
D 4.0 m/s 3.2 2.8 2.4 3.2 5.3 4.8 3.6 3.4 28.7
D 8.0 m/s 3.3 3.8 2.3 2.2 7.2 10.2 7.7 4.6 41.2
F 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 10.3 10.9 8.3 9.0 17.4 18.9 14.2 11.0 100.0
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 1.4 1.7 1.5 2.0 2.9 1.9 1.4 1.3 14.3
D 4.0 m/s 1.9 2.8 2.8 4.0 5.8 3.9 2.7 2.1 25.9
D 8.0 m/s 1.1 2.5 2.1 2.3 6.0 5.4 4.0 2.2 25.6
F 1.5 m/s 2.2 3.6 2.6 2.6 4.5 2.9 1.8 1.7 21.8
F 4.0 m/s 0.9 2.2 1.9 1.7 2.6 1.5 1.1 0.7 12.5
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 7.6 12.8 10.9 12.5 21.8 15.5 11.0 7.9 100.0
Model aspects 4.32
Rotterdam
346-015 2.17 0.84 2.26 1.88 0.00 0.00 7.16
016-045 1.97 0.84 1.62 1.42 0.00 0.00 5.85
046-075 2.86 0.85 2.13 2.23 0.00 0.00 8.07
076-105 2.91 0.84 2.02 1.89 0.00 0.00 7.66
106-135 1.58 0.52 1.40 0.93 0.00 0.00 4.43
136-165 1.31 0.88 1.61 0.81 0.00 0.00 4.60
166-195 1.66 1.19 3.26 2.44 0.00 0.00 8.54
196-225 1.64 1.08 3.76 4.86 0.00 0.00 11.34
226-255 2.04 1.31 3.86 7.11 0.00 0.00 14.33
256-285 2.75 1.36 4.09 4.38 0.00 0.00 12.57
286-315 2.40 0.87 2.74 2.88 0.00 0.00 8.90
316-345 1.22 0.61 2.01 2.72 0.00 0.00 6.57
Total 24.50 11.19 30.76 33.55 0.00 0.00 100.00
346-015 0.00 1.19 1.13 0.46 0.54 2.44 5.76
016-045 0.00 1.20 1.30 0.61 0.77 2.66 6.53
046-075 0.00 1.17 2.26 1.67 1.52 2.96 9.58
076-105 0.00 1.22 1.83 1.01 1.20 2.26 7.51
106-135 0.00 0.79 1.30 0.53 0.71 1.42 4.75
136-165 0.00 1.19 2.08 0.80 0.74 1.50 6.31
166-195 0.00 1.55 3.75 2.37 1.15 2.10 10.91
196-225 0.00 1.49 3.62 4.79 1.26 2.50 13.65
226-255 0.00 1.93 3.81 4.76 1.15 3.30 14.96
256-285 0.00 1.66 2.26 1.96 0.88 2.24 8.99
286-315 0.00 0.94 1.51 1.78 0.53 1.45 6.20
316-345 0.00 0.86 1.23 1.13 0.41 1.22 4.86
Total 0.00 15.19 26.06 21.87 10.85 26.04 100.00
Rotterdam
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 3.1 4.3 3.0 2.1 2.5 3.4 3.8 2.3 24.5
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 1.3 1.3 0.9 1.5 1.7 2.0 1.6 1.0 11.2
D 4.0 m/s 2.8 3.1 2.4 3.2 5.4 5.9 4.8 3.1 30.8
D 8.0 m/s 2.4 3.2 1.9 2.0 6.1 9.3 5.1 3.7 33.6
F 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 9.4 11.9 8.3 8.9 15.6 20.6 15.2 10.2 100.0
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 1.8 1.8 1.4 2.0 2.3 2.8 1.8 1.5 15.2
D 4.0 m/s 1.9 3.2 2.2 4.0 5.5 4.9 2.6 1.8 26.1
D 8.0 m/s 0.8 2.2 1.0 2.0 6.0 5.7 2.8 1.4 21.9
F 1.5 m/s 3.9 4.1 2.5 2.6 3.5 4.4 2.6 2.4 26.0
F 4.0 m/s 1.0 2.1 1.3 1.3 1.8 1.6 1.0 0.7 10.8
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 9.4 13.3 8.5 11.8 19.1 19.5 10.7 7.7 100.0
Model aspects 4.33
Schiphol
346-015 1.25 0.62 1.84 2.63 0.00 0.00 6.33
016-045 1.23 0.45 1.50 2.44 0.00 0.00 5.62
046-075 2.09 0.62 2.36 4.12 0.00 0.00 9.18
076-105 2.01 0.69 1.86 1.88 0.00 0.00 6.45
106-135 1.32 0.54 1.35 0.95 0.00 0.00 4.15
136-165 1.30 0.76 2.00 1.56 0.00 0.00 5.62
166-195 1.49 0.94 2.85 3.04 0.00 0.00 8.33
196-225 1.19 0.83 3.24 6.26 0.00 0.00 11.51
226-255 1.23 0.78 2.62 9.44 0.00 0.00 14.07
256-285 1.58 0.75 3.01 7.52 0.00 0.00 12.86
286-315 1.21 0.61 2.02 4.46 0.00 0.00 8.31
316-345 1.23 0.60 1.93 3.82 0.00 0.00 7.58
Total 17.12 8.17 26.59 48.12 0.00 0.00 100.00
346-015 0.00 0.83 1.60 1.03 0.83 1.87 6.15
016-045 0.00 0.55 1.24 1.33 0.69 1.04 4.84
046-075 0.00 0.75 2.15 3.11 1.21 1.27 8.49
076-105 0.00 0.90 2.42 2.20 1.63 1.53 8.68
106-135 0.00 0.86 1.60 0.67 0.83 1.36 5.32
136-165 0.00 1.14 2.74 1.81 1.27 1.61 8.57
166-195 0.00 1.51 3.76 2.99 1.31 2.10 11.66
196-225 0.00 1.19 4.14 5.99 1.38 1.36 14.06
226-255 0.00 1.24 2.66 5.28 1.01 1.75 11.94
256-285 0.00 0.96 1.77 3.60 0.67 1.26 8.26
286-315 0.00 0.73 1.35 2.36 0.49 1.03 5.96
316-345 0.00 0.86 1.65 1.48 0.62 1.46 6.06
Total 0.00 11.52 27.07 31.85 11.91 17.65 100.00
Schiphol
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 1.9 3.1 2.3 2.0 1.9 2.0 2.0 1.9 17.1
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 0.8 1.0 0.9 1.2 1.3 1.2 1.0 0.9 8.2
D 4.0 m/s 2.4 3.3 2.3 3.4 4.7 4.1 3.5 2.8 26.6
D 8.0 m/s 3.8 5.1 1.9 3.1 7.8 13.2 8.2 5.1 48.1
F 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 8.8 12.4 7.4 9.8 15.7 20.5 14.7 10.7 100.0
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 1.0 1.2 1.3 1.9 1.9 1.7 1.2 1.3 11.5
D 4.0 m/s 2.0 3.4 2.8 4.6 6.0 3.5 2.2 2.4 27.1
D 8.0 m/s 1.8 4.2 1.8 3.3 7.5 7.1 4.2 2.0 31.9
F 1.5 m/s 2.0 2.0 2.1 2.7 2.4 2.4 1.7 2.4 17.7
F 4.0 m/s 1.1 2.0 1.6 1.9 2.0 1.3 0.8 1.0 11.9
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 7.9 12.8 9.7 14.4 19.9 16.1 10.1 9.1 100.0
Model aspects 4.34
Soesterberg
346-015 1.99 1.49 2.59 1.53 0.00 0.00 7.60
016-045 3.75 1.74 2.86 1.44 0.00 0.00 9.79
046-075 2.16 1.18 1.67 1.02 0.00 0.00 6.03
076-105 2.33 1.11 1.61 1.20 0.00 0.00 6.25
106-135 1.62 0.98 1.35 0.48 0.00 0.00 4.43
136-165 1.33 1.34 1.76 0.57 0.00 0.00 5.01
166-195 1.51 2.03 3.01 1.19 0.00 0.00 7.74
196-225 1.65 2.45 5.37 3.53 0.00 0.00 13.00
226-255 1.39 1.59 4.70 5.15 0.00 0.00 12.83
256-285 1.51 1.58 3.82 4.84 0.00 0.00 11.76
286-315 1.64 1.25 3.99 2.73 0.00 0.00 9.62
316-345 1.04 1.13 2.16 1.64 0.00 0.00 5.96
Total 21.93 17.85 34.91 25.32 0.00 0.00 100.00
346-015 0.00 1.42 0.98 0.25 0.37 2.24 5.26
016-045 0.00 2.24 1.98 0.52 1.26 4.07 10.07
046-075 0.00 1.44 1.67 0.67 1.13 2.96 7.87
076-105 0.00 1.50 1.66 0.67 1.41 3.27 8.50
106-135 0.00 1.39 0.97 0.21 0.52 2.58 5.67
136-165 0.00 2.00 1.77 0.59 0.54 3.08 7.97
166-195 0.00 3.13 2.72 1.05 0.75 3.57 11.23
196-225 0.00 3.01 4.27 2.67 0.97 3.02 13.93
226-255 0.00 2.04 3.53 3.33 0.74 1.82 11.46
256-285 0.00 1.85 2.15 1.83 0.62 1.90 8.34
286-315 0.00 1.31 1.24 0.68 0.35 1.64 5.22
316-345 0.00 1.17 1.12 0.42 0.21 1.55 4.48
Total 0.00 22.49 24.07 12.88 8.88 31.69 100.00
Soesterberg
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 4.7 3.3 2.8 2.1 2.4 2.1 2.4 2.0 21.9
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 2.5 1.7 1.5 2.4 3.5 2.4 2.0 1.9 17.9
D 4.0 m/s 4.2 2.5 2.2 3.3 6.9 6.6 5.9 3.5 34.9
D 8.0 m/s 2.2 1.6 1.1 1.2 4.1 7.6 5.2 2.4 25.3
F 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 13.6 9.1 7.6 8.9 16.9 18.7 15.5 9.8 100.0
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 3.0 2.2 2.1 3.6 4.6 3.0 2.2 1.9 22.5
D 4.0 m/s 2.5 2.5 1.8 3.1 5.6 4.6 2.3 1.6 24.1
D 8.0 m/s 0.6 1.0 0.5 1.1 3.2 4.2 1.6 0.5 12.9
F 1.5 m/s 5.2 4.6 4.2 4.9 4.8 2.8 2.6 2.7 31.7
F 4.0 m/s 1.4 1.8 1.2 0.9 1.3 1.0 0.7 0.4 8.9
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 12.7 12.1 9.9 13.6 19.5 15.6 9.4 7.1 100.0
Model aspects 4.35
Twente
346-015 1.75 1.38 1.60 0.70 0.00 0.00 5.43
016-045 2.38 1.38 1.64 0.36 0.00 0.00 5.77
046-075 3.19 1.46 2.07 0.86 0.00 0.00 7.59
076-105 3.36 1.50 1.85 0.81 0.00 0.00 7.52
106-135 2.45 1.45 1.29 0.25 0.00 0.00 5.43
136-165 1.67 1.30 1.11 0.20 0.00 0.00 4.29
166-195 1.80 1.63 2.93 1.26 0.00 0.00 7.63
196-225 2.56 2.72 6.86 5.12 0.00 0.00 17.25
226-255 1.97 2.05 5.53 4.90 0.00 0.00 14.45
256-285 1.36 1.51 3.22 3.30 0.00 0.00 9.38
286-315 1.46 1.41 3.02 2.72 0.00 0.00 8.60
316-345 1.63 1.48 2.26 1.30 0.00 0.00 6.67
Total 25.59 19.25 33.38 21.78 0.00 0.00 100.00
346-015 0.00 1.07 0.69 0.18 0.27 1.46 3.66
016-045 0.00 1.48 1.16 0.14 0.61 2.66 6.04
046-075 0.00 1.81 2.00 0.54 1.67 3.39 9.41
076-105 0.00 1.73 1.86 0.61 1.94 3.24 9.39
106-135 0.00 1.71 1.18 0.16 1.25 2.88 7.18
136-165 0.00 1.60 1.15 0.22 0.60 2.49 6.06
166-195 0.00 2.26 3.21 1.13 1.47 3.03 11.10
196-225 0.00 3.19 5.98 4.32 1.73 3.35 18.57
226-255 0.00 2.21 4.09 3.40 0.99 1.99 12.68
256-285 0.00 1.46 2.06 1.57 0.52 1.51 7.12
286-315 0.00 1.38 1.30 0.84 0.35 1.37 5.24
316-345 0.00 1.10 0.88 0.28 0.17 1.12 3.56
Total 0.00 21.03 25.56 13.37 11.56 28.48 100.00
Twente
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 3.3 4.9 4.1 2.6 3.5 2.7 2.1 2.5 25.6
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 2.1 2.2 2.2 2.1 3.5 2.8 2.2 2.2 19.3
D 4.0 m/s 2.4 3.0 2.2 2.6 8.3 7.1 4.6 3.1 33.4
D 8.0 m/s 0.7 1.3 0.6 0.8 5.7 6.5 4.4 1.7 21.8
F 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 8.5 11.3 9.2 8.1 21.1 19.1 13.3 9.4 100.0
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 2.0 2.7 2.6 2.7 4.3 2.9 2.1 1.6 21.0
D 4.0 m/s 1.5 2.9 2.1 2.8 7.6 5.1 2.3 1.2 25.6
D 8.0 m/s 0.2 0.8 0.5 0.8 4.9 4.2 1.6 0.4 13.4
F 1.5 m/s 3.4 5.0 4.5 4.0 4.9 2.7 2.1 1.8 28.5
F 4.0 m/s 0.7 2.6 2.2 1.3 2.5 1.2 0.6 0.3 11.6
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 7.9 14.1 11.9 11.6 24.1 16.2 8.8 5.4 100.0
Model aspects 4.36
Valkenburg
346-015 1.93 0.65 2.40 4.32 0.00 0.00 9.30
016-045 1.26 0.75 1.59 1.61 0.00 0.00 5.20
046-075 1.93 0.81 2.01 2.87 0.00 0.00 7.62
076-105 1.89 0.72 1.79 1.99 0.00 0.00 6.39
106-135 1.16 0.51 1.26 1.39 0.00 0.00 4.32
136-165 1.44 0.78 1.76 1.57 0.00 0.00 5.56
166-195 1.32 0.96 2.12 2.43 0.00 0.00 6.84
196-225 0.76 0.85 2.74 5.24 0.00 0.00 9.59
226-255 1.00 0.79 3.01 9.86 0.00 0.00 14.66
256-285 2.13 0.99 3.94 6.77 0.00 0.00 13.83
286-315 1.71 0.76 2.38 3.96 0.00 0.00 8.81
316-345 1.58 0.65 2.11 3.55 0.00 0.00 7.89
Total 18.11 9.23 27.10 45.57 0.00 0.00 100.00
346-015 0.00 0.70 1.43 1.63 0.52 1.59 5.86
016-045 0.00 1.33 1.87 1.13 1.14 3.42 8.88
046-075 0.00 1.21 2.07 1.92 1.30 2.72 9.21
076-105 0.00 0.89 1.77 1.66 1.02 2.03 7.36
106-135 0.00 0.61 1.10 0.87 0.38 0.90 3.86
136-165 0.00 1.43 1.95 1.27 0.76 2.69 8.09
166-195 0.00 1.67 2.30 2.26 0.68 3.28 10.20
196-225 0.00 1.30 2.92 5.10 0.71 2.17 12.21
226-255 0.00 1.03 2.67 6.60 0.59 1.56 12.45
256-285 0.00 0.96 2.32 4.45 0.59 1.19 9.52
286-315 0.00 0.62 1.51 3.24 0.35 0.90 6.62
316-345 0.00 0.58 1.42 2.51 0.36 0.88 5.74
Total 0.00 12.34 23.33 32.63 8.38 23.32 100.00
Valkenburg
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 2.2 2.9 2.1 2.1 1.4 2.1 2.8 2.5 18.1
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 1.1 1.2 0.9 1.3 1.3 1.3 1.3 1.0 9.2
D 4.0 m/s 2.8 2.9 2.2 2.8 3.8 5.0 4.4 3.3 27.1
D 8.0 m/s 3.8 3.9 2.4 2.8 6.5 13.2 7.3 5.7 45.6
F 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 9.9 10.8 7.5 9.0 13.0 21.6 15.7 12.5 100.0
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 1.7 1.7 1.1 2.3 2.1 1.5 1.1 0.9 12.3
D 4.0 m/s 2.6 3.0 2.0 3.1 4.1 3.8 2.7 2.1 23.3
D 8.0 m/s 1.9 2.7 1.7 2.4 6.2 8.8 5.5 3.3 32.6
F 1.5 m/s 4.2 3.7 1.9 4.3 3.8 2.2 1.5 1.7 23.3
F 4.0 m/s 1.4 1.8 0.9 1.1 1.1 0.9 0.6 0.6 8.4
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 11.8 12.9 7.5 13.2 17.3 17.2 11.4 8.7 100.0
Model aspects 4.37
Vlissingen
346-015 2.10 0.64 2.44 2.80 0.00 0.00 7.98
016-045 2.31 0.76 2.16 2.23 0.00 0.00 7.46
046-075 1.89 0.58 1.86 2.62 0.00 0.00 6.95
076-105 2.28 0.54 1.52 1.49 0.00 0.00 5.82
106-135 1.91 0.58 1.41 0.99 0.00 0.00 4.89
136-165 1.23 0.50 1.36 1.18 0.00 0.00 4.28
166-195 1.19 0.51 2.22 3.99 0.00 0.00 7.91
196-225 1.19 0.54 2.60 6.72 0.00 0.00 11.04
226-255 1.71 0.66 2.47 9.42 0.00 0.00 14.26
256-285 2.84 0.68 3.56 8.42 0.00 0.00 15.50
286-315 1.36 0.56 2.02 2.93 0.00 0.00 6.86
316-345 1.57 0.63 2.19 2.68 0.00 0.00 7.07
Total 21.56 7.18 25.80 45.46 0.00 0.00 100.00
346-015 0.00 0.86 1.57 0.86 0.89 1.62 5.79
016-045 0.00 1.19 2.24 1.47 2.02 2.45 9.37
046-075 0.00 1.02 2.61 2.05 2.12 1.77 9.56
076-105 0.00 0.62 1.55 1.49 0.98 1.22 5.86
106-135 0.00 0.48 1.34 1.09 0.60 0.82 4.34
136-165 0.00 0.54 2.04 1.68 0.73 0.78 5.76
166-195 0.00 0.66 3.10 5.00 0.83 0.81 10.39
196-225 0.00 0.68 3.92 8.58 1.15 1.04 15.38
226-255 0.00 0.58 2.80 7.94 0.92 0.83 13.07
256-285 0.00 0.61 1.72 3.90 0.51 0.81 7.54
286-315 0.00 0.72 1.69 2.63 0.70 1.16 6.89
316-345 0.00 0.94 1.65 1.11 0.69 1.67 6.06
Total 0.00 8.89 26.24 37.79 12.12 14.96 100.00
Vlissingen
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 3.4 3.0 3.0 1.8 1.8 3.1 2.8 2.6 21.6
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 1.1 0.9 0.8 0.8 0.8 1.0 0.9 1.0 7.2
D 4.0 m/s 3.4 2.6 2.2 2.5 3.7 4.2 3.8 3.4 25.8
D 8.0 m/s 3.6 3.4 1.7 3.2 8.7 13.6 7.1 4.1 45.5
F 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 11.4 9.9 7.8 8.2 15.0 22.0 14.6 11.1 100.0
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 1.6 1.3 0.8 0.9 1.0 0.9 1.0 1.4 8.9
D 4.0 m/s 3.0 3.4 2.1 3.6 5.5 3.7 2.6 2.4 26.2
D 8.0 m/s 1.9 2.8 1.8 4.2 11.1 9.9 4.6 1.5 37.8
F 1.5 m/s 3.3 2.4 1.4 1.2 1.4 1.2 1.6 2.5 15.0
F 4.0 m/s 2.5 2.6 1.1 1.1 1.6 1.2 1.0 1.1 12.1
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 12.3 12.5 7.3 11.0 20.6 16.8 10.7 8.9 100.0
Model aspects 4.38
Volkel
346-015 2.11 1.38 1.91 0.94 0.00 0.00 6.34
016-045 2.23 1.22 1.65 1.06 0.00 0.00 6.15
046-075 3.02 1.08 1.95 2.03 0.00 0.00 8.08
076-105 2.50 0.92 1.49 1.41 0.00 0.00 6.31
106-135 1.76 0.76 1.18 0.79 0.00 0.00 4.49
136-165 1.50 1.04 1.45 0.95 0.00 0.00 4.94
166-195 1.60 1.61 2.56 1.91 0.00 0.00 7.68
196-225 2.12 2.18 4.35 4.79 0.00 0.00 13.43
226-255 2.45 2.37 5.90 6.24 0.00 0.00 16.95
256-285 2.00 2.11 4.24 3.97 0.00 0.00 12.33
286-315 1.59 1.48 2.70 1.94 0.00 0.00 7.71
316-345 1.32 1.21 1.94 1.10 0.00 0.00 5.58
Total 24.21 17.36 31.32 27.11 0.00 0.00 100.00
346-015 0.00 1.30 0.86 0.27 0.57 2.88 5.87
016-045 0.00 1.35 1.32 0.61 0.82 3.27 7.37
046-075 0.00 1.20 1.80 1.28 1.41 2.96 8.65
076-105 0.00 1.17 1.45 0.82 1.00 2.52 6.95
106-135 0.00 1.02 0.96 0.35 0.50 1.84 4.67
136-165 0.00 1.26 1.53 0.76 0.57 1.93 6.04
166-195 0.00 2.16 2.59 1.48 0.92 2.41 9.57
196-225 0.00 2.48 4.08 3.72 1.42 3.29 14.98
226-255 0.00 2.61 4.61 4.15 1.43 2.85 15.65
256-285 0.00 1.97 2.42 1.91 0.86 2.73 9.89
286-315 0.00 1.60 1.30 0.59 0.42 2.25 6.15
316-345 0.00 1.14 0.72 0.21 0.31 1.84 4.22
Total 0.00 19.24 23.64 16.13 10.22 30.77 100.00
Volkel
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 3.3 4.3 3.0 2.3 2.9 3.5 2.6 2.4 24.2
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 1.9 1.5 1.2 1.8 3.0 3.4 2.5 1.9 17.4
D 4.0 m/s 2.6 2.7 1.9 2.7 5.6 8.0 4.8 2.9 31.3
D 8.0 m/s 1.5 2.7 1.5 1.9 5.7 8.2 3.9 1.6 27.1
F 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 9.3 11.2 7.6 8.8 17.3 23.1 13.9 8.7 100.0
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 2.0 1.8 1.6 2.3 3.6 3.6 2.6 1.8 19.2
D 4.0 m/s 1.8 2.5 1.7 2.8 5.4 5.8 2.5 1.2 23.6
D 8.0 m/s 0.7 1.7 0.8 1.5 4.5 5.1 1.5 0.3 16.1
F 1.5 m/s 4.7 4.2 3.1 3.1 4.5 4.2 3.6 3.3 30.8
F 4.0 m/s 1.1 1.9 1.0 1.0 1.9 1.9 0.8 0.6 10.2
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 10.3 12.1 8.1 10.8 19.8 20.6 11.1 7.2 100.0
Model aspects 4.39
Woensdrecht
346-015 1.36 1.03 1.88 0.79 0.00 0.00 5.06
016-045 2.09 1.07 2.62 1.34 0.00 0.00 7.11
046-075 3.29 1.21 2.44 1.94 0.00 0.00 8.88
076-105 3.32 1.32 1.78 0.98 0.00 0.00 7.39
106-135 1.01 0.91 0.78 0.18 0.00 0.00 2.88
136-165 1.03 1.39 1.06 0.15 0.00 0.00 3.63
166-195 1.46 2.08 3.15 1.19 0.00 0.00 7.88
196-225 2.16 2.71 7.08 4.06 0.00 0.00 16.01
226-255 1.83 1.99 5.39 5.55 0.00 0.00 14.76
256-285 2.36 1.64 3.63 4.72 0.00 0.00 12.34
286-315 2.17 1.46 3.10 1.86 0.00 0.00 8.59
316-345 1.18 1.05 2.14 1.09 0.00 0.00 5.47
Total 23.24 17.87 35.05 23.83 0.00 0.00 100.00
346-015 0.00 1.16 0.72 0.14 0.28 1.95 4.25
016-045 0.00 1.51 1.55 0.64 1.11 2.93 7.74
046-075 0.00 1.67 2.41 1.38 1.88 4.10 11.44
076-105 0.00 1.81 1.20 0.47 0.95 4.04 8.46
106-135 0.00 1.70 0.58 0.09 0.19 2.30 4.87
136-165 0.00 1.88 0.81 0.08 0.23 2.44 5.43
166-195 0.00 3.02 3.00 1.18 0.76 3.25 11.21
196-225 0.00 3.62 5.80 3.21 1.78 4.00 18.40
226-255 0.00 2.36 4.46 3.18 1.05 2.44 13.48
256-285 0.00 1.14 1.55 1.68 0.40 1.21 5.97
286-315 0.00 1.17 1.20 0.72 0.38 1.38 4.85
316-345 0.00 1.22 0.84 0.25 0.18 1.42 3.91
Total 0.00 22.24 24.12 13.02 9.19 31.44 100.00
Woensdrecht
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 2.8 4.9 2.7 1.8 2.9 3.0 3.3 1.9 23.2
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 1.6 1.9 1.6 2.4 3.8 2.8 2.3 1.6 17.9
D 4.0 m/s 3.6 3.3 1.7 2.6 8.7 7.2 4.9 3.1 35.1
D 8.0 m/s 1.7 2.4 0.7 0.7 4.7 7.9 4.2 1.5 23.8
F 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 9.6 12.6 6.6 7.6 20.0 20.9 14.8 8.0 100.0
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 2.1 2.6 2.6 3.4 5.1 2.9 1.7 1.8 22.2
D 4.0 m/s 1.9 3.0 1.2 2.3 7.3 5.2 2.0 1.2 24.1
D 8.0 m/s 0.7 1.6 0.3 0.7 3.8 4.0 1.6 0.3 13.0
F 1.5 m/s 3.9 6.1 4.3 4.1 5.6 3.0 2.0 2.4 31.4
F 4.0 m/s 1.2 2.4 0.7 0.6 2.2 1.2 0.6 0.3 9.2
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 9.9 15.7 9.1 11.0 24.0 16.5 7.8 6.0 100.0
Model aspects 4.40
Ypenburg
346-015 1.71 0.95 2.41 2.58 0.00 0.00 7.65
016-045 1.65 1.00 1.81 1.33 0.00 0.00 5.78
046-075 2.72 1.14 2.36 2.82 0.00 0.00 9.04
076-105 2.20 0.98 1.66 1.82 0.00 0.00 6.65
106-135 1.74 0.77 1.32 1.06 0.00 0.00 4.90
136-165 0.88 0.73 1.10 0.73 0.00 0.00 3.43
166-195 0.77 0.95 2.19 2.08 0.00 0.00 5.98
196-225 1.10 1.14 3.41 5.11 0.00 0.00 10.75
226-255 1.51 1.26 3.54 7.41 0.00 0.00 13.73
256-285 2.34 1.40 3.76 8.39 0.00 0.00 15.90
286-315 1.42 0.76 2.38 3.35 0.00 0.00 7.90
316-345 1.53 0.88 2.46 3.42 0.00 0.00 8.29
Total 19.56 11.95 28.40 40.09 0.00 0.00 100.00
346-015 0.00 1.22 1.42 0.77 0.67 2.34 6.42
016-045 0.00 1.62 1.67 0.90 1.36 3.05 8.60
046-075 0.00 1.56 2.41 2.00 1.72 3.69 11.37
076-105 0.00 1.22 1.40 0.97 0.75 2.19 6.52
106-135 0.00 1.05 1.15 0.49 0.45 1.49 4.63
136-165 0.00 1.05 1.16 0.73 0.35 1.29 4.57
166-195 0.00 1.30 2.34 2.06 0.61 1.30 7.60
196-225 0.00 1.25 4.16 5.07 1.16 1.64 13.27
226-255 0.00 1.84 3.31 4.88 1.03 2.69 13.74
256-285 0.00 1.49 2.36 4.25 0.56 1.85 10.50
286-315 0.00 0.76 1.54 2.68 0.42 0.97 6.36
316-345 0.00 1.03 1.78 1.86 0.50 1.25 6.42
Total 0.00 15.38 24.69 26.62 9.58 23.74 100.00
Ypenburg
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 2.5 3.8 2.8 1.3 1.5 2.7 2.6 2.4 19.6
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 1.5 1.6 1.3 1.2 1.6 2.0 1.5 1.4 11.9
D 4.0 m/s 3.0 3.2 2.2 2.2 4.5 5.4 4.3 3.7 28.4
D 8.0 m/s 2.6 3.7 2.0 1.8 6.2 11.6 7.5 4.7 40.1
F 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 9.6 12.4 8.2 6.4 13.7 21.7 15.9 12.1 100.0
B 1.5 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 4.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
B 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
D 1.5 m/s 2.2 2.2 1.7 1.7 1.9 2.6 1.5 1.6 15.4
D 4.0 m/s 2.4 3.1 1.8 2.3 5.3 4.5 2.7 2.5 24.7
D 8.0 m/s 1.3 2.5 1.0 1.8 6.1 7.0 4.8 2.2 26.6
F 1.5 m/s 4.2 4.8 2.6 1.9 2.3 3.6 1.9 2.4 23.7
F 4.0 m/s 1.7 2.1 0.8 0.7 1.5 1.3 0.7 0.8 9.6
F 8.0 m/s 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Total 11.8 14.6 7.9 8.4 17.1 19.0 11.6 9.6 100.0
Model aspects 4.41
4.C.1 Outflow models
Table 4.1 gives the correlation between the LOC and the model to be used, making a number of
choices:
The modelling of a totally ruptured vessel filled with compressed gas and (non-boiling) liquid
is in agreement with the information provided in the Yellow Book, Section 2.5.1.
The continuous releases of vessels, tanks and transport units cover several LOCs, namely the
10-min release and the holes with effective diameters of 10 mm as well as the largest
connection in the liquid phase. It is selected to model the release as a hole in the vessel wall
with a sharp orifice (C
d
= 0.62), as this is the most appropriate way to describe the rupture of
small instrumentation pipes in the vessel wall, and the loss of containment due to external
impact and a welding rupture. A rounded orifice with C
d
= 0.95 to 0.99 might be more
appropriate for modelling the continuous release of the largest connection in the liquid phase
for transport units However, for simplicity only one type of orifice with corresponding C
d
values has been selected. The value of C
d
is in agreement with the recommendations in the
Yellow Book and is almost equal to the value given in [IPO], C
d
= 0.6.
The release from a leak in a pipeline is modelled as a stationary one. The pressure upstream is
assumed to be kept constant due to the presence of a large upstream vessel or of
compressors/pumps.
The default value for the wall roughness, , is the same as for commercial steel, as listed in the
Yellow Book (Table 2.2). The value is intermediate, falling between the value for bronze, lead
and glass ( = 1.5 m) and the value for cast iron ( = 250 m).
The models in the Yellow Book for release from the liquid phase assumes that pure liquid
flows out of the vessel, i.e. flashing of liquid takes place outside the vessel. Some integrated
risk models allow flashing in the orifice. The assumption of liquid flow without flashing in the
orifice is consistent with the value assumed for C
d
, i.e. C
d
= 0.62. In some circumstances the
modelling of flashing in the orifice can be more appropriate. The modelling of flashing in the
orifice should be well motivated.
The location of the hole is determined by the specific design of the tank or vessel. Important
aspects to be considered are e.g. the position of the connections to the tank, the position of
relief devices and the distribution of dangerous substances within a column. It might not be
worthwhile to use a number of release positions for each part of an installation. In generic
studies, no specific information is available. Hence, a default value is selected for the location
of the hole. The default value should be near the geographical level at which people live, in
order not to underestimate the risk. The value selected, i.e. one metre above ground,
corresponds to the reference height for the calculation of effects.
The release duration is limited to a maximum of 30 minutes:
For a continuous release of flammables, the maximum mass in the flammable cloud, i.e. the
mass within the explosion limits, is usually reached within a few minutes. After this initial
time period, the mass added to the flammable cloud is compensated by the dispersion of
mass out of the flammable cloud. Therefore the release duration of flammables can be
limited to 30 minutes.
For a release of toxics, the maximum exposure duration is set at 30 minutes (see Section
5.2.2). Therefore the release duration of toxics can be limited to 30 minutes.
Model aspects 4.42
The direction of release is determined by the specific design of the tank or vessel. However, if
no information is available, a default value is selected. The horizontal direction is in line with
current practice. Models assume that release is parallel to the wind direction, so this direction
is selected.
Underground transport pipelines form an exception. As underground transport pipelines are
embedded in soil, a vertical release is selected as default.
The conditions for obstructed flow are taken from [IPO]. A calculation method for the length
of the free jet L
j
is added. The length of the free jet, L
j
, is derived from the velocity of the gas
in the free jet, according to equation 4.76 of the Yellow Book [CPR14E]:
u
c
(s) / u
0
= C
u
b
0
/ s, (4.C.1)
where:
u
c
velocity in the jet (m s
-1
)
u
0
-1
)
C
u
empirical constant, equal to 12 (-)
b
0
source radius (m)
s co-ordinate along jet axis (m)
This equation implies that the release is in a uniform quiescent atmosphere.
The length of the free jet is calculated using the criterion that the velocity at the end of the free
jet regime is comparable to the ambient wind velocity. The length of the free jet, L
j
, is then
defined as the position, s, where the velocity in the jet, u
c
(s), is equal to the ambient wind
velocity u
air
:
L
j
= C
u
u
0
b
0
/ u
air
(4.C.2)
where:
C
u
empirical constant, equal to 12 (-)
u
0
-1
)
b
0
source radius (m)
u
air
ambient wind velocity (m s
-1
)
For a simple expression for the length of the free jet, independent of the meteorological
situation, the ambient wind velocity is set to an average of 5 m s
-1
.
The modelling of an obstructed free jet, using a reduction of the impulse of the jet with a factor
4, is copied from [DNV98].
4.C.2 Blocking systems
The default values for the operation of the blocking system are taken from [IPO]. In [IPO], a
distinction is made between failure upon demand of the blocking system, and failure upon
demand of the blocking valves. The distinction is not well described for the three different types
of blocking systems. Therefore the failure upon demand is described here as the failure of the
Model aspects 4.43
blocking system as a whole. The closing time of the systems is based on the following
considerations [IPO]:
The closing time for an automatic blocking system is based on a fully automated gas detection
system. The closing time of 2 min consists of:
- 30 s for the gas to reach the detector;
- 30 s for the closing signal from the detector to reach the closing valve;
- 1 min to close the valves.
The closing time for a remote controlled blocking system is based on an automated gas
detection system. The closing time of 10 min consists of:
- 30 s for the gas to reach the detector;
- 30 s for the alert signal from the detector to reach the control room;
- 7 min to validate the signal
- 2 min to close the valves.
The closing time for a hand-operated blocking system is based on a automated gas detection
system. The closing time of 30 min consists of:
- 30 sec for the gas to reach the detector;
- 30 sec for the alert signal from the detector to reach the control room;
- 7 min to validate the signal;
- 15 min for the operator to go to the blocking valve and to make use of personal
protective equipment;
- 7 min to remove the security locks and to close the valves.
It should be noted that the effect of a hand operated blocking system does not appear in the
QRA, as the maximum duration of the outflow is equal to 30 min.
4.C.3 Repression systems
The application of a repression system is in line with [IPO], where it is explicitly stated that the
effect of repression systems should be demonstrated and quantified. The repression systems
considered in this section reduce the source term following a LOC. Other repression, like a
sprinkler installation on a pressurised storage tank of LPG or the cooling of tank wagons in the
vicinity of a fire systems, affect the probability of a LOC. The effects of these types of
repression systems are discounted in the failure frequencies.
4.C.4 Pool evaporation
The default values in Table 4.2 and Table 4.3 are taken from the Yellow Book. Since the models
to calculate pool evaporation are described in terms of circular pools, an effective pool radius is
defined.
4.C.5 Time-varying release
The conversion of a time-varying release into one single release segment is included since a
number of models are yet not able to handle multiple time segments. For flammables, a high
release rate is selected as the effects of flammable gases are usually determined by the mass
released in the first few minutes of the release. Hence, the release conditions of the first 20% of
the mass is assumed to be decisive. For toxics, the effects are determined by the dose received in
the total exposure duration. Hence, the outflow conditions should be more or less averaged over
the total release duration. However, a weighing towards the higher release rates is necessary since
Model aspects 4.44
the lethality of most toxics is more than linear in the concentration. Consequently, the conditions
of the mass released in the second out of five time segments are assumed to be representative.
The modelling of the dispersion as independent steady-state releases is selected as a more
conservative approach.
4.C.6 Modelling vapour cloud dispersion
The deposition of material can be modelled if data are available. However, often the models used
do not consider deposition. Therefore deposition processes do not have to be considered in the
dispersion calculation.
The default values for the roughness length are taken from the Yellow Book.
The averaging time for flammables is short, as indicated in the Yellow Book. The selection of the
value is based on currently used values. The averaging time for toxics should be comparable to the
exposure time and consequently to the release duration. If a large number of scenarios have to be
processed, it is not worthwhile calculating each scenario with an individual averaging time. The
averaging time for toxics is based on an exposure time of 10 minutes. This value is somewhat
arbitrarily selected, i.e. between short release with exposure time in the order of 30 s - 60 s, and a
long release with exposure time in the order of 30 min.
4.C.7 Release inside a building
The calculation of the release of substances from a building is based on the assumption that the
concentration in the room where the release takes place is (almost) immediately in equilibrium
with the new source term. Where this assumption is not correct, the concentration in the
ventilation air, C
vent
, has to be calculated with the differential equation:
V dC
vent
/dt = Q
in
C
vent
F, (4.C.3)
with V being the volume of the room in m
3
and Q
in
the (time-dependent) source term inside the
building. The release to the atmosphere, Q
out
, is equal to Q
out
= C
vent
F.
The influence of the building wake was recently described in [Wi98]. The equation for the
concentration in the recirculation zone is also described in [IPO]. In [IPO], the length of the
recirculation zone is determined by the height of the building. In [NM86], the length of the
recirculation zone is determined by the minimum of either the height or the width of the building.
The formulation in [NM86] is used here. The default value K = 1, being the value for the yearly
averaged concentration used for buildings in general is applied here.
4.C.8 Fires and plume rise
The Yellow Book uses the Briggs formula for plume rise. The formulas in the Yellow Book
were originally derived for releases from conventional stacks and are probably not applicable to
open fires. Analogous to the CPR-15 method, the concentration of toxic substances at ground
level is therefore assumed to be low due to the plume rise and the dilution in the atmosphere.
Lethal effects are therefore not to be expected. However, if a model is available to calculate the
Model aspects 4.45
concentration at ground level for open fires, it is recommended to use it to calculate the plume
trajectory.
4.C.9 Ignition
The data on direct ignition and the probability of a BLEVE are taken from [IPO].
For stationary installations, it is assumed that 70% of the instantaneous releases with direct
ignition is caused by heating due to a pool fire underneath the tank or heat radiation from a fire
nearby. The result is a BLEVE and a fire ball. Since the excess heat of the fire results in an excess
pressure in the tank, the total inventory of the tank is incorporated in the BLEVE.
It is assumed that 30% of the instantaneous releases with direct ignition is due to other causes.
Since no excess heat is available, a vapour cloud and a liquid pool may be formed. Direct ignition
of the vapour cloud results in a flash fire and explosion (see Section 4.8), whereas direct ignition
of the liquid pool results in a pool fire.
Recently, the fraction that will remain airborne has been reviewed [VITO97]. The review
indicates that the fraction that will remain airborne is probably larger than twice the flash fraction
and recommends using the relationships given in Table 4.8 for instantaneous releases. Hence, the
relationships are copied here.
The pressure at failure of the vessel should be equal to 1.21 the opening pressure of the relief
device or, if no relief device is present, should be equal to the test pressure of the vessel.
The effect of a fireball depends on the fraction of the generated heat radiated by the fireball
[CPR14E]. The fraction of the heat radiated is a number between 0.2 and 0.4 and is a function of
the vapour pressure of the flammable material inside the vessel at failure. Consequently, the burst
pressure of the vessel is an important factor.
Table 7.1 of the Yellow Book gives characteristic pressure and temperature at failure [CPR14E].
If failure is due to causes like corrosion of the vessel, erosion of the vessel, material defect,
external impact or fatigue of vessel, the pressure at failure is the storage or working pressure. If
the failure cause is external fire, the pressure at failure is equal to 1.21 the opening pressure of
the safety valve. Causes like overfilling or overheating in combination with failure of the safety
valve results in failure at the design pressure a safety factor (= usually 2.5).
The effect of the pressure on the effect distances of a BLEVE are shown in Figures 4.C.1 and
4.C.2 for storage of 100 tonnes propane and butane, respectively. The pressures at failure shown
are:
the storage pressure at ambient temperature, 282 K;
1.21 the relief pressure of the safety valve; the relief pressure is set equal to the saturated
vapour pressure at 308 K;
the test pressure; the test pressure is calculated by multiplying the overpressure at 308 K +
1.7 bar by a factor of 1.4. The relief pressure should always be less than the test pressure
[CPR8-3];
the pressure corresponding with the maximum fraction of heat radiated, 0.4; this pressure is
equal to 34 bar or higher.
Model aspects 4.46
Figure 4.C.1 The probability of death as a function of the distance for a BLEVE of a storage
tank containing 100 tonne propane. Indicated are the storage pressure at ambient
temperature (6.2 bar), 1.21 the relief pressure of the safety valve (15 bar), the
test pressure (18 bar) and the curve corresponding with maximum fraction of heat
radiated (> 34 bar).
Figure 4.C.2 The probability of death as a function distance for a BLEVE of a storage tank
containing 100 tonne butane. Indicated are the storage pressure at ambient temperature (1.4 bar),
1.21 the relief pressure of the safety valve (4.0 bar), the test pressure (5.6 bar) and the curve
corresponding with maximum fraction of heat radiated (> 34 bar).
Frequently, the storage or working pressure is used in the effect calculations of a BLEVE.
However, it is postulated that a BLEVE is caused by weakening of the tank vessel wall due to a
Model aspects 4.47
pool fire underneath a tank or due to a fire nearby. Hence, the pressure at failure is set equal to
1.21 the relief pressure of the safety valve. If no safety valve is present, the pressure at failure
is set equal to the test pressure of the tank.
For delayed ignition, in [IPO], a distinction is made between stationary installations and
transport units at an establishment. For stationary units, delayed ignition is always equal to (1
direct ignition), whereas for transport units non-ignition is defined as a probability. There is no
argument for this distinction.
The free field calculation results in risk contours that are independent of the environment. As a
conservative approach, the delayed ignition should be modelled to give the maximum effects. In
general, this is an iterative process since the mass in the cloud decreases as the distance travelled
increases. To facilitate the calculations, it is decided that delayed ignition occurs at maximum
cloud extent. A definition of the maximum cloud extent would be the maximum amount of mass
within the LFL contour. However, using this definition, an instantaneous release would ignite
immediately at the release location. It is therefore decided to use as a definition of maximum cloud
extent: the maximum area of the LFL cloud footprint.
In principle, an Individual Risk calculation should be done with calculation method B, the free
field calculation. However, on special occasions it is allowed to use calculation method A. Since
criteria to use calculation method A have not been defined yet, the decision has to be made by the
competent authorities.
It has been decided to model substances that are both toxic and flammable as flammables using
the probability of direct ignition, and otherwise as toxics. However, substances like ammonia are
usually modelled as purely toxics. Therefore substances with low reactivity are to be modelled as
purely toxics.
To determine the probability of ignition, the approach described in [DNV96, AM94] is used (see
Appendix 4.A). It should be noted that the figures are not completely reliable and some care has
to be taken. Therefore, using a time-independent probability of ignition is not excluded, in which
case the value for the passage of the cloud is arbitrarily set to 1 min.
[IPO] gives some figures for the (time-independent) probability of ignition. A comparison with
the figures used in [AM94] is given in Table 4.C.1; here the length of a road element is set equal
to 100 m, and the velocity of a vehicle to 50 km/h.
Table 4.C.1 Ignition probability as given in [IPO] and [AM94]
Source IPO [AM94]
industrial site 0.9 0.9
process installations 0.5 0 - 0.9
road, N < 50 veh. per hour 0.5 0 - 0.1
road, N > 50 veh. per hour 1 0.1 - 1
Model aspects 4.48
4.C.10 Effects of ignition of a vapour cloud
Following the delayed ignition of an unconfined vapour cloud, one event occurs which has the
characteristics of both a flash fire and an explosion. The range of possible events can be assumed
to be divided in two separate classes, namely a pure flash fire and a pure explosion. A reasonable
estimation seems to be a 60% 40% division [DNV96]. However, in the study called LPG-
Integraal a different division is used, namely 30% 70% division for a flash fire and explosion,
respectively [TNO83]. As these figures are uncertain, one of these values is selected arbitrarily.
The effect distances to the peak side on overpressure contours of 0.1 barg and 0.3 barg, R
0.1 barg
and R
0.3 barg
respectively, are calculated using the Multi-Energy Method [CPR14E]. The
combustion energy scaled distances are derived from Figure 5.8A in [CPR14E], using the highest
value of the blast strength, 10. The values of the combustion energy scaled distances
corresponding to 0.1 barg and 0.3 barg, r
0.1 barg
and r
0.3 barg
respectively, are estimated as r
0.1 barg
=
3 and r
0.3 barg
= 1.5.
The results of the Multi Energy Method using the highest value of the blast strength are
comparable to a TNT equivalent method with a TNT-equivalence factor of 20% in the
overpressure range of 10 to 100 kPa [CPR14E]. Previously, distances to overpressure contours
are calculated as R
0.3
= 0.03 ( E)
1/3
and R
0.1
= 0.06 ( E)
1/3
[CPR14]. The TNT-
equivalence factor, , was set equal to 10%. Using a default value of 0.08 for the fraction of the
mass in obstructed regions would therefore reproduce the distances used previously.
4.C.11 Rupture of vessels
The blast waves and high velocity fragments following the rupture of a pressurized vessel can
lead to lethal effects outside the plant area. However, these effects are not considered in the QRA
for the following reasons:
The effect of the physical blast wave is expected not to be important at larger distances, i.e.
outside the plant area, relative to the toxic or flammable effects.
The probability of being hit by a fragment outside the plant area is very low.
Therefore these effects will not to be considered in the QRA.
4.C.12 Meteorological data
In contrast to [IPO] it was decided not to define the exact wind speed, but to use the categories
low, medium and high. In this way the transformation of data can be avoided if an available wind
speed distribution does not exactly match the classification used in [IPO].
In the Manual [IPO] it was decided to use the closest meteorological station. However, the
closest meteorological station is not necessarily the most representative. Especially for a site near
the coast, it is better to use a meteorological station near the coast instead of an inland station.
The frequency distributions of a number of meteorological stations are listed in Appendix 4.B.
The data are compiled from [KNMI72], are based on the Pasquill classification of stability using
routine meteorological observations as wind speed, cloud cover and time of day and are used in
QRAs up till now. Recently, a new classification method of stability is proposed, resulting in a
Model aspects 4.49
reduced frequency of neutral conditions [NNM98]. It is strongly recommended to use frequency
distributions based on the new classification method when statistical data are available. However,
long-term statistical data may not yet be available for the meteorological stations. The use of
average data over the Netherlands or the use of data of weather stations relatively far away may
lead to discrepancies between the situation to be modelled and the data used.
The values in this manual for the meteorological parameters differ from the values listed in [IPO]
in that they are more appropriate as yearly averaged values. The values are derived from
[KNMI92]:
Ambient air temperature equal to the yearly averaged value, 9.3 C
soil/bund temperature set equal to the ambient air temperature
Water temperature set equal to the ambient air temperature
Ambient pressure equal to the yearly averaged value, 1015.1 hPa
Humidity equal to the yearly averaged value, 83%
solar radiation flux equal to the average global radiation in a year (364,584 J cm
-2
)
divided by the time period of one year
In reality, a small difference, in the order of 1 C, exists between the average soil temperature and
the average air temperature. As the difference is small, it has been decided to set the various
temperatures equal to one another.
Exposure and Damage 5.1
5. MODELLING EXPOSURE AND DAMAGE
5.1 Introduction
Determination of the exposure and effects follows the release and dispersion of a dangerous
substance in the environment. Since regulation is based on the probability of death, only lethal
effects are relevant. This chapter will describe the calculation methods to determine the
probability of death given the exposure and the fraction of the population for whom exposure is
fatal. The following two parameters are used throughout this chapter to express the lethal effects:
the probability of death, P
E
, indicating the probability of an individual dying from exposure.
The individual is assumed to be outdoors and to be unprotected. This parameter, P
E
, is to be
used in the calculation of the Individual Risk contours.
the fraction of the population dying, F
E
, indicating the fraction of the population dying at a
certain location due to a given exposure. At least part of the population is protected by
staying indoors and wearing protective clothing. For this reason, two values are used, F
E, in
and F
E, out
, to denote the respective fractions of the population dying indoors and outdoors.
The parameters, F
E, in
and F
E, out
, are to be used in the calculation of the Societal Risk.
Probit functions are used to calculate the probability of death due to toxic substances and
exposure to heat radiation at a given exposure. The use of probit functions is described in Section
5.2.1. The effects of toxic substances, fires and explosions are described in the Sections 5.2.2,
5.2.3 and 5.2.4, respectively. Section 5.3 presents the guidelines to calculate the population
present in certain surroundings.
5.2 Damage modelling
5.2.1 Probit functions
The calculation of the Individual Risk and the Societal Risk involves calculating the probability of
death of a person at a given exposure. The probability of death is calculated using probit
functions. The relation between the probability of an effect, P, and the corresponding probit, Pr,
is given by:
(5.1)
where:
(5.2)
The relation between the probability of an effect, P, and the corresponding probit, Pr, is also
given in Table 5.1.
The relation between the probability of an effect and the exposure usually results in a sigmoid
curve. The sigmoid curve is replaced with a straight line if the probit is used instead of the
probability, as shown in Figure 5.1.
Table 5.1 The probit, Pr, as a function of the probability, P
P 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
0 - 2.67 2.95 3.12 3.25 3.36 3.45 3.52 3.59 3.66
0.1 3.72 3.77 3.82 3.87 3.92 3.96 4.01 4.05 4.08 4.12
0.2 4.16 4.19 4.23 4.26 4.29 4.33 4.36 4.39 4.42 4.45
0.3 4.48 4.50 4.53 4.56 4.59 4.61 4.64 4.67 4.69 4.72
0.4 4.75 4.77 4.80 4.82 4.85 4.87 4.90 4.92 4.95 4.97
0.5 5.00 5.03 5.05 5.08 5.10 5.13 5.15 5.18 5.20 5.23
0.6 5.25 5.28 5.31 5.33 5.36 5.39 5.41 5.44 5.47 5.50
0.7 5.52 5.55 5.58 5.61 5.64 5.67 5.71 5.74 5.77 5.81
0.8 5.84 5.88 5.92 5.95 5.99 6.04 6.08 6.13 6.18 6.23
0.9 6.28 6.34 6.41 6.48 6.55 6.64 6.75 6.88 7.05 7.33
Figure 5.1 The probability, P, and the probit, Pr, as a function of exposure to ammonia. The
exposure is represented by ln (D), with D the toxic dose (see Section 5.2.2). The figure
shows how the sigmoid curve is replaced with a straight line if the probit is used.
5.2.2 Toxic exposure
Figure 5.2 Calculation of the probability of death, P
E
, and the respective fractions of the
population dying indoors and outdoors, F
E,in
and F
E,out
, due to exposure to a toxic
cloud. The function f(a,b,n;C,t) is the probit function for exposure to toxic substances
(toxics).
The probability of death due to exposure to a toxic cloud, P
E
, and the fraction of people indoors
and outdoors dying, F
E,in
and F
E,out
, is given in Figure 5.2. The probability of death, P
E
, is
calculated with the use of a probit function and relation 5.1 in Section 5.2.1. The probit function
for death due to toxic exposure is given by:
Pr = a + b x 1n (C
n
x t), (5.3)
with:
Pr probit corresponding to the probability of death (-)
a, b, n constants describing the toxicity of a substance (-)
C concentration (mg m
-3
)
t exposure time (minutes)
Notes:
1. The value of the constant, a, depends on the dimensions of the concentration, C, and the
exposure time, t. The dimensions of the concentration and exposure time must correspond
to the value of the constant, a, in the probit function.
2. The probit is a function of the toxic dose to an individual. The toxic dose, D, is equal to
D = C
n
t if the concentration, C, is constant during the time of exposure, t. If the
concentration is not constant in time, the toxic dose is calculated as D = C
n
dt; the probit,
Pr, should be calculated correspondingly.
3. The exposure time, t, is limited to a maximum of 30 minutes, starting from the arrival of the
cloud. The arrival of the cloud can be defined as the moment when the probability of death,
P
E
, exceeds 1%.
4. Staying indoors reduces the toxic dose since the concentration indoors is lower than the
concentration outdoors during cloud passage. This effect is accounted for by a generic
factor 0.1 in the fraction of people dying indoors. Instead of using the generic factor 0.1,
one may use the ventilation rate of dwellings to calculate the dose indoors and the fraction
of people dying indoors. The method is described in [CPR16]. It should be noted that the
dose indoors strongly depends on the passage time of the cloud, the ventilation rate during
the passage of the cloud and the ventilation after the passage of the cloud.
The passage time of the cloud varies with distance from the source and is different for
each LOC. Hence, the dose indoors has to be calculated for each LOC and at each
distance.
The ventilation rate depends strongly on parameters like the type and age of the
dwellings, weather conditions and the opening and closing of windows. If no specific
information is available, a ventilation rate of 1 h
-1
and no adsorption should be used.
People are not expected to know when the cloud has passed. Consequently, the
ventilation rate after the passage of the cloud is equal to the ventilation rate during the
passage of the cloud and the lag-time between cloud passage and the start of total
ventilation by opening the windows is high, i.e. minimally 0.5 hours.
The maximum exposure time indoors is 30 minutes. The lag-time between cloud passage
and the start of total ventilation by opening the windows is therefore maximally 0.5
hours.
5. The values for the constants describing the toxicity of a substance, a, b and n, are given for
a number of substances in Table 5.2. The toxic constants of these and a number of other
substances are also given in the database of substances [RIVM99].
The probit function of a toxic substance not listed in Table 5.2 or in the database of
substances [RIVM99] has to be determined on the basis of acute toxicity data of animals.
The procedure to derive the constants a, b and n from toxicity data is described in
[CPR16]. However, expert advice is needed when using toxicity data to derive the probit
function.
Table 5.2 Values for the constants describing the toxicity of a substance, a, b and n. The values
are valid for the probit function with the concentration, C ( mg m
-3
) and the exposure
time, t (min).
Substance a b n
Acrolein 4.1 1 1
Acrylonitrile 8.6 1 1.3
Allylalcohol 11.7 1 2
Ammonia 15.6 1 2
Azinphos-methyl 4.8 1 2
Bromine 12.4 1 2
Carbon monoxide 7.4 1 1
Chlorine 6.35 0.5 2.75
Ethylene oxide 6.8 1 1
Hydrogen chloride 37.3 3.69 1
Hydrogen cyanide 9.8 1 2.4
Hydrogen fluoride 8.4 1 1.5
Hydrogen sulfide 11.5 1 1.9
Methyl bromide 7.3 1 1.1
Methyl isocyanate 1.2 1 0.7
Nitrogen dioxide 18.6 1 3.7
Parathion 6.6 1 2
Phosgene 10.6 2 1
Phosphamidon 2.8 1 0.7
Phosphine 6.8 1 2
Sulphur dioxide 19.2 1 2.4
Tetraethyllead 9.8 1 2
5.2.3 Fire
E
, where the respective fractions of the
population dying indoors and outdoors are F
E,in
and F
E,out
on exposure to a flash
fire.
Figure 5.4. Calculation of the probability of death, P
E
, where the respective fractions of the
population dying indoors and outdoors are F
E,in
and F
E,out
, for exposure to a
BLEVE, pool fire and jet fire. The probit function for heat radiation is f(Q,t).
The probability of death due to a flash fire, P
E
, and the respective fractions of people dying
indoors and outdoors, F
E,in
and F
E,out
, is given in Figure 5.3. The probability of death due to a
BLEVE, jet fire or pool fire, P
E
, and the respective fractions of people dying indoors and
outdoors, F
E,in
and F
E,out
, is given in Figure 5.4. The probability of death due to the exposure to
heat radiation is calculated with the use of a probit function and relation 5.1 in Section 5.2.1. The
probit function for death due to heat radiation is given by:
Pr = 36.38 + 2.56 x 1n (Q
4/3
x t), (5.4)
with:
Pr probit corresponding to the probability of death (-)
Q heat radiation (W m
-2
)
t exposure time (s)
Notes:
1. The probit function for death due to heat radiation is currently under review. The function
may be modified in a new edition of the Green Book [CPR16].
2. The flame envelope of a flash fire is equal to the LFL contour at the time of ignition.
3. The exposure time, t, is equal to the duration of the fire. However, the exposure time is
limited to a maximum of 20 s.
4. It is assumed that people indoors are protected from heat radiation until the building
catches fire. The threshold for the ignition of buildings is set at 35 kW m
-2
. If the building is
set on fire, all the people inside the building are assumed to die. Hence, F
E,in
= 1 if the heat
radiation, Q, exceeds 35 kW m
-2
and F
E,in
= 0 if the heat radiation, Q, is less than 35 kW m
-
2
.
5. For the Societal Risk calculation, it is assumed that people outdoors are protected from
heat radiation by clothing until it catches fire. The protection of clothing reduces the
number of people dying by a factor of 0.14 compared to no protection of clothing. The
threshold for the ignition of clothing is set at 35 kW m
-2
and people die if clothing catches
fire at this threshold. Hence, F
E,out
= 1 if the heat radiation, Q, exceeds 35 kW m
-2
and F
E,out
= 0.14 P
E
if the heat radiation, Q, is less than 35 kW m
-2
.
5.2.4 Pressure effects for a vapour cloud explosion
E
, and the respective fractions of the
population dying indoors and outdoors, F
E,in
and F
E,out
from exposure to a
blast/explosion.
The probability of death due to an explosion, P
E
, and the respective fractions of people dying
indoors and outdoors, F
E,in
and F
E,out
, are given in Figure 5.5. It should be noted that the values
given here are only applicable to a vapour cloud explosion. The values are not applicable to the
detonation of explosives due to differences in the duration of the blast.
5.3 Population
5.3.1 Survey of the population present
The presence of the population is important to the calculation of both the Societal Risk and the
ignition probability. Therefore the presence of population in the environment should be
surveyed. The following rules apply:
1. The present situation should be used to determine the population present. The population
in future residential areas for which approved spatial plans exist should be taken into
account. The population density in these areas has to be calculated on the basis of these
plans. If no actual information is available, default values can be used for the population
density in new residential areas [CPR16]. The accuracy of the information needed also
depends on the purpose of the QRA.
2. The population in recreational areas should be taken into account. If the type of recreation
depends on the season, discrete values for the population density should be used for
different times of the year. If large groups of people are present during short time periods,
as in stadiums, discrete values for the population density should be used for different time
periods. However, if the fraction of the time that a large group of people is present is very
low, the use of discrete values can be omitted. This can, for example, happen if a stadium is
used only for short periods of time throughout the year. As a guideline, if the product of
the summed frequency of the relevant scenarios and the fraction of time that a large group
of people is present is less than 10
-9
per year, the presence of the large group can be
ignored.
3. Current legislation has to be taken into account to decide whether people present in e.g.
industrial areas, offices, hospitals and schools, on (other) industrial sites and motorways
should be taken into account
a
. Inclusion or exclusion of groups of people also depends on
the purpose of the QRA.
4. The presence of population varies with time, as people travel out of the area to work,
attend schools and the like. Therefore different values have to be used for the population
during daytime and night-time. The following rules are applied to determine the presence of
the population:
Daytime refers to the period from 8:00 to 18:30 MET, night-time to the period 18:30 -
8:00 MET.
In residential areas the fraction of the population present during daytime is 0.7. The
fraction of the population present during night-time is 1.0. If schools and/or working
places are present in the residential area, the presence of people in these locations
should be taken into account.
In industrial areas the fraction of the population present during daytime is equal to 1.0.
If work is also done in night shifts, the fraction of the population present during night-
time is equal to 0.2, otherwise the fraction is set equal to 0.
In recreational areas the population present during daytime and night-time depends on
the type of recreation.
In principle, the population data used should be as detailed as possible. In practice, it can be
difficult to collect population data on a very detailed level. There are a number of sources
available to supply information on the population, like the municipalities and the provinces.
Information can also be obtained on a commercial basis. Geographical Information Systems can
be useful in processing population data. The following types of data, in decreasing order of detail,
may be available:
1. The location of each house, e.g. by counting the number of houses on a map. The
population density is derived assuming a density of 2.4 persons per house. Please note that
one building on a map, e.g. an apartment building, may consist of several houses.
2. The central point of a group of houses, e.g. from databases listing the centre of gravity of all
houses having the same postal code. The population density is derived assuming a density
of 2.4 persons per house. However, care should be taken if large apartment buildings are
present, in which case, the centre of gravity of the postal code can be some distance from
the location of the apartments.
a
The regulation to include or exclude groups of people is given in the new AMvB.
3. If no information on the population density is available at all, e.g. if spatial plans are only
known roughly, default values for the population density can be used. Default values for
various types of areas are listed in [CPR16].
As the population in a residential area is based on an average population density, people on local
roads are already included in the average population density.
5.3.2 Fraction indoors and outdoors
In the calculation of the Societal Risk, it is assumed that at least part of the population is
protected by staying indoors and wearing protective clothing. Since different values are used for
the fractions of the population dying indoors and outdoors, the respective fractions of the
population present indoors and outdoors, f
pop, in
and f
pop, out
, have to be set. Default parameter
values are given in Table 5.3. The values are valid for residential and industrial areas unless other
information is available. For recreational areas, the type of recreation determines the fraction of
the population indoors and outdoors.
Table 5.3 Fraction of the population present indoors (f
pop, in
) and outdoors (f
pop,out
) for daytime
and night-time, where daytime refers to 8:00 - 18:30 MET and night-time to 18:30 -
8:00 MET
f
pop, in
f
pop, out
daytime 0.93 0.07
night-time 0.99 0.01
Appendix 5.A COMMENTARY
5.A.1 General
The relation between the probability of an effect, P, and the corresponding probit, Pr, in Table 5.1
is taken from [CPR16].
5.A.2 Toxic exposure
For toxic substances, the probit function for the probability of death is described in [CPR16].
The values in Table 5.2 for the toxic constants a, b and n are taken from [KO 24-2], with the
exception of ammonia and phosgene. The probit function of ammonia is taken from [KO 59] and
of phosgene from [KO 86]. For a number of substances two different probit functions are listed
in [KO 24-2], one function assuming n = 1 and another assuming n = 2. In [CPR16] it is
recommended to use the value n = 2 in cases where n is not known. Hence only the probit
function with n = 2 is given here.
The exposure time, t, is limited to a maximum of 30 minutes, in agreement with [IPO].
A maximum exposure time needs to be set because the probit function shows that for continuous
releases of small amounts of substances all humans will eventually suffer death if the exposure
time is long enough. As this is obviously not to be expected, a maximum exposure time is defined.
Why exactly 30 minutes is taken as the maximum exposure time is not known. This value is
possibly based on time needed for evacuation and/or the time lapse before first aid is provided.
The exposure time indoors may be longer than the passage time of the cloud, depending on the
ventilation rate of the building. The exposure time indoors is limited to 30 minutes and takes into
account the opening of windows after the passage of the cloud.
The concentration of toxic substances indoors can be calculated using a ventilation model for the
building. However, the calculation is quite elaborate and the concentration indoors depends,
among other aspects, on the specific meteorological conditions and the ventilation regime.
Furthermore, the concentration inside the building is not uniform but varies from room to room,
depending on the ventilation within the building [TNO98a]. Despite these difficulties, a
ventilation model can be used to estimate the concentration inside a building. The use of a
ventilation model is described in [CPR16]. Default parameter values are no adsorption and a
ventilation rate of 1 h
-1
(in line with the data in [CPR16]). The ventilation rate is assumed to be
unchanged after the passage of the cloud and the maximum time of exposure to be 30 minutes. If
the passage time of the cloud is relatively short, the dose reduction can then be estimated using a
lag-time of 0.5 h between the passage of the cloud and the start of total ventilation.
In practice, while ventilation models are rarely used in a QRA, a generic reduction factor is used.
[IPO] applies a reduction factor of 0.1 to the total population present, indoors and outdoors.
The fraction of people dying is presented separately for the population indoors and outdoors.
The population outdoors is not protected, so no reduction factor is applied to F
E, out
.
People indoors are protected. A reduction factor of 0.1 is applied to F
E, in
. During daytime, 7%
is assumed to be outdoors. A reduction factor of 0.1 in daytime for the fraction of people
dying indoors corresponds to a factor of 0.16 for the total population. During night-time,
almost all humans (99%) are assumed to be indoors (see Section 5.3.2), and only 1% outdoors.
At night-time, a reduction factor of 0.1 for the fraction of people dying indoors leads to
almost the same results as a factor of 0.1 for the total population. Furthermore, at night-time
the effect distances are largest since stable weather conditions with low wind speeds prevail.
Hence, it was decided to apply a reduction factor of 0.1 to F
E, in
.
5.A.3 Fire
The probability of death due to a flash fire, BLEVE, jet fire or pool fire are are derived from
[CPR16, IPO].
An alternative calculation method to the probit function is described in [KO 20-2], a damage area
being defined. For example, the damage area of a flash fire is the flame envelope and the damage
area of a BLEVE is the area in which heat radiation exceeds a level of 12.5 kW m
-2
. In the damage
area, the probability of death is defined separately for people remaining indoors or outdoors. It is
advised not to use this method anymore, but rather the method described here [IPO, CPR16].
Inside the flame envelope the probability of death is equal to one due to the high level of heat
radiation and the ignition of clothing and buildings. Outside the flame envelope of a flash fire the
heat radiation is assumed to be low and the probability of death equal to zero. The effects of a
BLEVE, jet fire and pool fire outside the flame envelope are determined by the heat radiation.
The flame envelope of a flash fire is equal to the LFL contour, i.e. expansion of the cloud during
combustion is not considered.
The maximum exposure time is set equal to 20 s. It is assumed that people can flee to a safe place
within 20 s. It should be noted that previously it was assumed that people can reach a safe place
more easily in a built-up area and consequently the maximum exposure time in a built-up area
was reduced to 10 sec in the Societal Risk calculation [IPO]. However, it was decided to set the
maximum exposure time equal to 20 s everywhere to facilitate the calculations at establishments
where both built-up and untilled areas exist in the surroundings.
In the Individual Risk calculation, the probability of death from heat radiation is calculated with
the probit function for an unprotected individual. The protective effect of clothing is not
considered in the Individual Risk calculation.
In the Societal Risk calculation, the protection of clothing is taken into account and the
probability of death due to heat radiation is calculated for an individual outdoors wearing clothing.
The protective effect of clothing is assumed to reduce the probability of death; a factor of 0.14 is
applied [CPR16]. However, if clothing ignites, the probability of death is equal to one. The
threshold for the ignition of clothing is equal to Q
2
t = 2.5 10
4
kW
2
m
-4
s [CPR16]. An
exposure time of 20 s results in a threshold of 35 kW m
-2
. As it is assumed that buildings are set
on fire at this level of heat radiation, this threshold for the ignition of clothing, 35 kW m
-2
, is used
in built-up areas as well.
1. In QRAs for establishments, the protective effect of buildings was not considered. A
reduction factor of 0.14 was applied to people indoors and outdoors [IPO].
2. In QRAs for marshalling yards, people died at heat radiation levels exceeding 40 kW m
-2
.
Below that level, people were considered safe inside buildings [SAVE97]. However, a
reduction factor of 0.14 for people outdoors was not applied.
The modelling described here assumes that people are safe inside a building until the building
catches fire. If the building is set on fire, all people inside are likely to succumb. Unfortunately,
there was little information found on the threshold of ignition for buildings. Information indicates
that [CPR16]:
glass breaks at 4 kW m
-2
the critical heat intensity, i.e. the radiation level where ignition occurs for long exposure times,
is in the range of 25 - 35 kW m
-2
for wood, textiles, fibreboard, hardboard and plastics. The
critical heat intensity is in the order of 10 - 15 kW m
-2
if ignition flames are present, e.g. fire
sparks.
ignition of upholstery occurs if the heat radiation, Q, and the exposure time, t, meets the
condition Q
2
t 2.5 - 4.5 10
4
kW
2
m
-4
s. At an exposure time, t = 20 s, the lowest value
corresponds to Q 35 kW m
-2
.
In view of this information, the threshold of the ignition of buildings is set to 35 kW m
-2
.
A sample calculation for the effects of a BLEVE of 100 tonne LPG is shown in Figure 5.A.1.
A large fraction of the people are usually assumed to be indoors (see Section 5.3.2). It is therefore
to be expected that for establishments the calculation method described here will result in a
reduction in the Societal Risk relative to the calculation method used previously. For marshalling
yards, the effect is assumed to be small. On one hand, the increase in the threshold radiation level
from 35 kW m
-2
to 40 kW m
-2
will increase the Societal Risk. On the other hand, the factor 0.14
for people outdoors will decrease the Societal Risk.
It should be noted that a BLEVE also leads to blast effects. These effects are currently not
included in the QRA calculation. The side-on peak overpressure of the blast wave can be
calculated with the method described in Chapter 7 of the Yellow Book [CPR14E]. TNO has
made a number of sample calculations to determine the blast effects relative to the heat radiation
effects [TNO98d]. Results are shown in Figure 5.A.2, where distances to various pressure and
radiation levels are shown as a function of tank inventory:
1. 0.03 bar overpressure, corresponding to the critical overpressure causing windows to break;
2. 0.1 bar overpressure, corresponding to 10% of the houses severely damaged and a
probability of death indoors equal to 0.025 (see Section 5.2.4);
3. radius of the fireball, corresponding to 100% lethality
4. heat radiation equal to 35 kW m
-2
, corresponding to the heat intensity at which people
indoors are fully protected.
The protective effect of buildings is considered explicitly in the Societal Risk calculation.
Previously, two different approaches were used:
Figure 5.A.1 Example calculation of a BLEVE of 100 tonne propane (burst pressure 15 bar).
The exposure time is set at the maximum value of 20 s. Indicated are heat
radiation, Q, probability of death, P
E
, fraction of people dying outdoors, F
E,out
, and
fraction of people dying indoors, F
E,in
.
Figure 5.A.2 Distances to various effect levels of a BLEVE as a function of tank inventory
[TNO98d]. Indicated effect levels are the radius of the fireball, heat radiation equal
to 35 kW m
-2
and overpressure levels of 0.1 bar and 0.03 bar.
The results show that for individual risk calculations, the pressure effects of a BLEVE can be
ignored since the fire ball effects are dominant compared to overpressure effects on unprotected
individuals. However, for societal risk calculations, caution is required. The results show that the
blast effects at the distance corresponding with a heat radiation equal to 35 kW m
-2
are between
0.1 bar and 0.03 bar overpressure. Consequently, breakage of windows is likely to occur, leading
to glass fragments and reduced protection. Furthermore, glass also partly transmits heat radiation,
and may also break due to intense heat radiation. Since nowadays large wall surfaces are made of
glass, especially in office buildings, it is questionable whether the assumption that people are
fully protected indoors at heat radiation levels of 35 kW m
-2
is valid.
The assumption that people indoors are fully protected if the heat intensity is less than
35 kW m
-2
is therefore based on the following:
people indoors are able to find shelter behind walls in a very short time
people indoors do not suffer lethal effects from being hit by glass fragments.
Note that this last assumption agrees with the effects of overpressure as described in Section
5.2.4, where no lethal effects are expected below 0.1 bar overpressure.
5.A.4 Pressure effects
The probability of death, P
E
, due to the exposure to pressure waves is calculated using three
different zones, depending on the peak overpressure, P
peak
. The values in Figure 5.5 are derived
from the values given in [IPO].
The origin of the value 0.025 for the fraction of people dying indoors is not well established. The
study LPG-Integraal indicates that: (1) about 10% of the houses outside the cloud and inside
the 0.1 barg contour are severely damaged and (2) about one out of eight persons present in a
severely damaged house is killed [TNO83]. This would suggest a factor of 0.1 0.125 = 0.0125
for the fraction of people suffering death indoors, i.e. a factor of two less.
5.A.5 Population
The values on the presence of population are mainly extracted from [CPR16].
[CPR16] gives a range for the fraction of the population present during daytime (0.3 0.7); the
maximum value is selected here.
The period corresponding with daytime is set at 8:00 - 18:30 MET. It corresponds to the period
used in the meteorological data.
The recommended number of persons per house is derived from the total population and number
of houses in the Netherlands [BR97].
The presence of large groups of people can be ignored if the product of the summed frequency of
the relevant scenarios and the fraction of time that a large group of people is present is less than
10
-9
per year. This criterion is derived from the presentation of the Societal Risk curve, since
frequencies below 10
-9
per year do not have to be shown in the Societal Risk curve (see Chapter
6).
The values of the respective fractions of the population present indoors and outdoors, f
pop, in
and
f
pop, out
, are also taken from [CPR16].
Calculation and presentation 6.1
6. CALCULATION AND PRESENTATION OF RESULTS
6.1 Introduction
The results of a QRA are the Individual Risk and the Societal Risk.
The Individual Risk represents the frequency of an individual dying due to loss of
containment events (LOCs). The individual is assumed to be unprotected and to be present
during the total exposure time, as indicated in Chapter 5. The Individual Risk is presented as
contour lines on a topographic map.
The Societal Risk represents the frequency of having an accident with N or more people being
killed simultaneously. The people involved are assumed to have some means of protection, as
indicated in Chapter 5. The Societal Risk is presented as an FN curve, where N is the number
of deaths and F the cumulative frequency of accidents with N or more deaths.
This chapter describes a calculation method for the Individual Risk and the Societal Risk,
followed by a presentation of the results.
6.2 Calculation of the Individual Risk and the Societal Risk
The methods for calculating the Individual Risk and the Societal Risk are described for toxic and
flammable substances. Section 6.2.1 describes the definition of the grid, Section 6.2.2 the
calculation method of the Individual Risk and Section 6.2.3, the method used to calculate the
Societal Risk. The definition of ignition events is described in Section 6.2.4. A crucial step is the
calculation of the probability of death and the fraction of people suffering death. The
computation of these factors is explained for toxic and flammable substances in Sections 6.2.5
and 6.2.6, respectively.
The calculation procedure described here is used in a number of computer programs. The
description does not cover all possible events but is mainly intended to illustrate the principles of
the calculation. Among other possible calculation methods is the one presented here, which
assumes that weather data are available in the form of frequency tables of weather classes and
wind directions (see Chapter 4). A different approach would be to use Monte Carlo simulations
of a large number of weather sequences.
Note:
1. The probability of death and the fraction of people suffering death should be calculated up
to the level of 1% lethality.
6.2.1 Definition of the grid
The calculation of the Individual Risk and the Societal Risk starts with the definition of a grid
over the area of interest, the calculation grid. The centre of a grid cell is called a grid point and the
Individual Risk is to be calculated at each grid point separately. The size of the grid cell should be
small enough not to influence the calculation results, i.e. the Individual Risk may not vary much
within a grid cell. As a guideline, if the effect distances of the significant scenarios are less than or
comparable to 300 m, the size of a grid cell should not be larger than 25 25 m. For effect
distances of the significant scenarios larger than 300 m, a grid cell of 100 100 m can be used. If
relevant, a combination can be used, i.e. a small grid cell in the calculation up to 300 m and a large
grid cell in the calculation starting from 300 m.
Next, the population within each grid cell has to be determined. The location of the population is
determined following the guidelines in Chapter 5. Each location of population (e.g. a house or the
central location of a group of houses) is assigned to a grid cell and the population is distributed
uniformly over the entire grid cell, i.e. within a grid cell a uniform population density is assumed.
It should be noted that a location may represent a large group of houses that extend over several
grid cells. In these cases, it is advised to distribute the population over a representative number of
grid cells.
Finally, a probability of ignition is assigned to each grid cell. All ignition sources in the grid cell are
combined into a single ignition source located at the centre of the grid cell.
6.2.2 Individual Risk calculation
The Individual Risk is calculated at each grid point separately. The procedure to determine the
Individual Risk at a single grid point is outlined in Figure 6.1. The frequency of an individual
dying is calculated at a grid point for each Loss of Containment event (LOC), each weather class,
each ignition event i (flammables only) and each wind direction separately. Next, the Individual
Risk at the grid point is determined by adding up all contributions.
The various steps in the procedure to calculate the Individual Risk, IR, at a grid point:
1. Select a LOC, S. The failure frequency of the LOC is given by f
S
(in y
-1
). The various LOCs
and the corresponding failure frequencies are described in Chapter 2.
2. Select a weather class, M, with probability P
M
(-). The various weather classes are
described in Section 4.10. Select a wind direction, , with conditional probability P (-).
The conditional probability, P ,, is the probability of obtaining wind direction, , given the
weather class, M. Often the product P
M
P is given, being the probability of obtaining
weather class, M, and wind direction, , simultaneously.
3. In case of the release of flammables, select an ignition event, i, with conditional probability
P
i
(-). The ignition events are described in Section 6.2.4.
4. Calculate the probability of death at the grid point, P
d
, given the LOC, S, the weather class,
M, the wind direction, , and ignition event, i (flammables). The calculation of P
d
is
elaborated in Section 6.2.5 for toxic substances and Section 6.2.6 for flammable substances.
The reference height for the calculation of effects is equal to one metre.
5. Calculate the contribution, IR
S, M, , i
of the LOC, S, the weather class, M, the wind
direction, , and ignition event, i, to the Individual Risk at the grid point:
(6.1)
Figure 6.1 Procedure to calculate the Individual Risk, IR, at a grid point.
6. Repeat the calculation steps 3 - 5 for all ignition events, 2 - 5 for all weather classes and
wind directions and 1 - 5 for all LOCs. The total Individual Risk, IR, at the grid point is
calculated as:
(6.2)
6.2.3 Societal Risk calculation
The procedure to determine the Societal Risk is outlined in Figure 6.2. For a single combination of
LOC, weather class, wind direction, and ignition event, the expected number of deaths is
calculated for each grid cell. Next, the expected number of deaths in all grid cells, N, is calculated
for each combination of LOC, weather class, wind direction and ignition event separately. Finally,
the cumulative frequency of having more than N deaths is determined.
The method to calculate the Societal Risk is outlined in Figure 6.2.
The various steps in the procedure to calculate the Societal Risk:
1. Select:
- an LOC, S, with frequency f
S
(y
-1
)
- a weather class, M, with probability P
M
(-)
- a wind direction, , with conditional probability P (-)
- ignition event, i, with conditional probability P
i
(-), flammables only
The probability of weather class, M, and wind direction, , occurring simultaneously is
given by P
M
P .
2. Select a grid cell. The number of people in the grid cell is N
cell
.
3. Calculate the fraction of deaths, F
d
, in the grid cell given the LOC, S, the weather class, M,
the wind direction, , and ignition event, i. This calculation is elaborated in Section 6.2.5 for
toxic substances and Section 6.2.6 for flammable substances. The reference height for the
calculation of effects is equal to one metre.
4. Calculate the expected number of deaths in the grid cell, N
S,M,
, given the LOC, S, the
weather class, M, the wind direction, , and ignition event, i.
N
S,M, ,i =
F
d
x N
cell
, (6.3)
The expected number of deaths in the grid cell is not necessarily a whole number.
5. Repeat steps 2 - 4 for all grid cells. Calculate the contribution of all grid cells to the total
number of deaths, N
S,M, ,i
, for the LOC, S, weather class, M, and wind direction, , and
ignition event, i.
N
S,M, ,i =
all grid cells

N
S,M, ,i
(6.4)
Figure 6.2 Procedure to calculate the Societal Risk.
6. Calculate the frequency, f
S,M, ,i
, of the combined LOC, S, weather class, M, wind direction,
, and ignition event, i.
f
S,M, ,i
= f
S
P
M
P P
i
(6.5)
7. Repeat the calculation steps 1 - 6 for all LOCs, weather classes and wind directions and
ignition events. The FN curve is now constructed by cumulating all frequencies f
S,M, ,i
for
which N
S,M, ,i
is greater than or equal to N:
FN = f
S,M, ,i
with N
S,M, ,i
N (6.6)
S,M, ,i
6.2.4 Definition of ignition events for flammable substances
A release of flammable substances results in various events depending on whether there is a direct
or delayed ignition. The situation is outlined in Figure 6.. The tree of events shows possible
outcomes following direct and delayed ignition for the time steps 0 - T, T - 2 T, .... ; T is
the time step used in the calculations. The possible outcomes of the event tree are:
BLEVE
jet fire and pool fire
flash fire
explosion
Figure 6.3 Tree of events for the release of flammable substances, using actual ignition sources.
Two different calculation methods can be used, depending on the manner delayed ignition is dealt
with, to calculate the various ignition events, i, and their conditional probability, P
i
(see Chapter
4). Calculation method A, using actual ignition sources, is described in Section 6.2.4.1. This
calculation method is used to determine the Societal Risk and, on special occasions, the Individual
Risk. Calculation method B, a free field calculation, is described in Section 6.2.4.2. This
calculation method is used to determine the Individual Risk.
6.2.4.1 Calculation with actual ignition sources (method A)
The method using actual ignition sources to calculate the ignition events i and their conditional
probability P
i
, given LOC, S, weather class, M, and wind direction, , is outlined in Figure 6.4.
Figure 6.4 Calculation of the ignition events, i, and the conditional probability for a release of
flammables using actual ignition sources.
The various steps are illustrated below:
1. Select LOC, S, weather class, M, and wind direction, ,
2. Calculate the probability of direct ignition. This defines the first ignition event, direct
ignition.
3. If the probability of direct ignition is not equal to one, calculate the characteristics of the
cloud after time step T.
4. If there is flammable mass within the LFL contour, determine the probability of delayed
ignition in time interval T. This probability is equal to (1 probability of direct ignition)
multiplied by the probability of delayed ignition, given the presence of the flammable
cloud. This defines the second ignition event, delayed ignition in time interval 0 - T.
5. Calculate the summed probability of all ignition events defined.
6. Repeat steps 3 - 5 as long as the summed probability of all ignition events defined is less
than one and as long as there is flammable mass within the LFL contour.
6.2.4.2 Free field calculation (method B)
The free field calculation uses actual ignition sources at the establishment only. If ignition does
not occur at the establishment, ignition occurs at maximum cloud extent, being defined as the
maximum area of the footprint of the LFL contour (see Chapter 4). The free field method to
calculate the ignition events i and their conditional probability P
i
, given LOC, S, weather class, M,
and wind direction, , is outlined in Figure 6.5.
Figure 6.5 Calculation of the ignition events, i, and the conditional probability for a release of
flammables using the free field method.
The various steps are illustrated below:
1. Select LOC, S, weather class, M, and wind direction, ,
2. Calculate the probability of direct ignition. This defines the first ignition event, direct
ignition.
3. If the probability of direct ignition is not equal to one, calculate the characteristics of the
cloud after a time step T.
4. If there is flammable mass within the LFL-contour, determine the probability of delayed
ignition in time interval T, using only the ignition sources at the establishment. This
probability is equal to (1 probability of direct ignition) multiplied by the probability of
delayed ignition, given the presence of the flammable cloud. This defines the second
ignition event, delayed ignition in time interval 0 - T.
5. Calculate the summed probability of all ignition events defined.
6. Repeat steps 3 - 5 as long as the summed probability of all ignition events is less than one,
as long as there is flammable mass within the LFL-contour and as long as the LFL-contour
of the cloud covers part of the area of the establishment.
7. If (1) the LFL-contour of the cloud is outside the boundary of the establishment, (2) the
summed probability of all ignition events is less than one and (3) there is flammable mass
within the LFL-contour, an additional ignition event is defined with probability (1
summed probability of all ignition events). Ignition takes place at maximum cloud extent,
defined as the maximum area of the footprint of the LFL-contour. The position of the cloud
and the flammable mass within the cloud will be determined iteratively.
6.2.5 Probability of death, P
d
, and fraction of deaths, F
d
, for toxic substances
A crucial step in the procedure to determine the Individual Risk and the Societal Risk is the
calculation of the probability of death at grid point, P
d
, and the fraction of deaths in the grid cell,
F
d
, given a Loss of Containment event, a weather class and a wind direction (see Figure 6.1 and
Figure 6.2). In this section the procedure to calculate these factors is described for a release of
toxic substances. It is assumed that the toxic clouds are limited in width and cover only one wind
sector. The situation where the cloud covers a large number of wind sectors is described in
Section 6.2.6, where the calculation procedure for a release of flammable substances is described.
The method to calculate the probability of death at a grid point, P
d
, and the fraction of deaths in a
grid cell, F
d
, given the LOC, S, the weather class, M, and the wind direction, , is outlined in
Figure 6.. Essentially, the toxic cloud at the position of the grid point is replaced with an effective
cloud having a uniform probability of death. Next, the probability of death at grid point, P
d
, is
calculated by multiplying the probability of death within the (effective) cloud by the probability
that the grid point is located inside the (effective) cloud. The various steps are illustrated below:
1. Calculate the distance, R, between the grid point and the source.
2. Calculate the concentration C(R, t) on the centre line at height h = 1 m for all times, t,
following the methods described in Chapter 4.
3. Calculate the probability of death, P
cl
(R), and the fraction of deaths, F
cl
(R), on the centre
line of the plume at distance R and at the reference height h = 1 m, according to the
relationships described in Section 5.2.2.
The fraction of deaths, F
cl
(R), is calculated as the sum of the fraction of deaths indoors,
F
E,in
, multiplied by the fraction of people indoors, f
pop,in
, and the fraction of deaths
outdoors, F
E,out
, multiplied by the fraction of people outdoors, f
pop,out
:
F
cl
= F
E,in
f
pop,in
+ F
E,out
f
pop,out
(6.7)
4. The probability of death decreases away from the centre line of the plume. Calculate the
probability of death, P(R,y), as a function of the distance to the centre line, y.
5. Calculate the effective width of the cloud, ECW, at distance R.
The toxic cloud is now replaced at distance R with an effective cloud, a uniform cloud with
a constant probability of death equal to the probability of death on the centre line of the
toxic cloud. The effective cloud width is determined by the standard that the number of
deaths in the toxic cloud is equal to the number of deaths in the effective cloud. The
effective cloud width can be calculated with the probability integral. The probability
integral, PI (R), is defined as the integral of the probability of death along the axis
perpendicular to the centre line of the plume:
(6.8)
with:
y distance to the centre line of the plume
P(R,y) probability of death at location (R,y).
In a numerical calculation, the boundaries of the integral can be taken as the distance
corresponding with 1% lethality.
d
, and the fraction of deaths at grid point, F
d
,
for toxic releases.
The effective cloud width, ECW, is then defined as:
(6.9)
6. Calculate the probability, P
ci
, that the grid point is covered by the effective cloud. In the
first approximation, the probability that the grid point is covered by the effective cloud is
given by:
(6.10)
if the grid point is situated in the wind sector and P
ci
(R) = 0 if the grid point is not situated
in the wind sector. The parameter n
ws
is equal to the number of wind sectors.
The validity of this equation is discussed in Appendix 6.A.
d
, as:
P
d
P
cl
P
ci
(6.11)
8. Calculate the fraction of deaths in the grid cell as:
F
d
F
cl
P
ci
(6.12)
A samplelateculation is given in Appendix 6.B.
6.2.6 Probability of death, P
d
, and fraction of deaths, F
d
, for flammables
In this section the procedure to calculate the probability of death at a grid point, P
d
, and the
fraction of deaths in the grid cell, F
d
, given a Loss of Containment event, S, a weather class, M, a
wind direction, , and an ignition event, i, is described for the release of flammable substances.
The impact area is assumed to be large, covering at least several wind sectors. In this way, the jet
fire, the pool fire, the flash fire, as well as the explosion, can be situated on the centre line of the
wind sector. The condition that the impact area will cover several wind sectors is, in general, valid
if the cloud is ignited nearby the source. The situation in which the impact area is small with
respect to the boundaries of the wind sector is described in Section 6.2.5, where the calculation
procedure for a release of toxic substances is described.
The various steps are outlined in Figure 6.7:
1. Calculate the flame envelope in the event of fire and the contours of 0.3 and 0.1 bar
overpressure in the event of a vapour cloud explosion.
2. Calculate the location of the grid point relative to the flame envelopes, overpressure
contours and the source of heat radiation.
3. Calculate the heat radiation, Q(x,y,t), at the grid point if the grid point is outside the flame
envelope in the event of BLEVE, and jet and pool fires. The calculation method is described
in the Yellow Book.
d
, following the methods described in
Chapter 5.
5. Calculate the fraction of deaths in the grid cell, F
d
, following the methods described in
Chapter 5. The fraction of deaths, F
d
, is calculated as the sum of the fraction of deaths
indoors, F
E,in
, multiplied by the fraction of people indoors, f
pop,in
, and the fraction of deaths
outdoors, F
E,out
, multiplied by the fraction of people outdoors, f
pop,out
:
F
d
= F
E,in
f
pop,in
+ F
E,out
f
pop,out
(6.13)
Note:
1. In the event of a flash fire, the probability of death changes abruptly from one to zero. To
have a more gradual change in the Individual Risk and Societal Risk, it is recommended to
set P
d
and F
d
equal to the fraction of the grid cell covered by the flash fire flame envelope if
the grid cell is partially located in the flame envelope. The same procedure can be applied
to the contours of 0.3 and 0.1 bar overpressure in the event of a vapour cloud explosion.
d
, and the fraction of deaths at a grid point,
F
d
, for a release of flammables.
6.3 Presentation of the results
The results of a QRA are the Individual Risk and the Societal Risk, both of which should be
presented clearly.
The Individual Risk must be presented as contour lines on a normal topographic map. The
Individual Risk contours with frequencies of 10
-4
, 10
-5
, 10
-6
, 10
-7
and 10
-8
per year must be
displayed, if in existence. The scale of the topographic map should be appropriate to
displaying the risk contours. It is advised to use one of the standard scales, i.e. 1 : 10,000,
1 : 25,000, 1: 50,000 or 1: 250,000. An illustration is shown in Figure 6.7.
The Societal Risk must be plotted as an FN-curve.
The x-axis of the FN-curve represents the number of deaths, N. The number of deaths is
placed on a logarithmic scale and the minimum value should be displayed as 1.
The y-axis of the FN-curve represents the cumulative frequency of the accidents, with the
number of deaths equal to N or more. The cumulative frequency is placed on a logarithmic
scale and the minimum value should be displayed as 10
-9
y
-1
.
An illustration is shown in Figur.
As well as the presentation of the Individual Risk and the Societal Risk, other results can be
presented, which will evaluate and judge the risks. Useful presentations can be figures of the
probability of death on the centre line, P
cl
, as a function of distance, x, for the decisive LOCs.
Figure 6.8 Presentation of the Individual Risk contours. Shown are the Individual Risk contours
10
-5
, 10
-6
, 10
-7
and 10
-8
y
-1
of a fictive plant .
Figure 6.8 Presentation of the societal risk curve. Indicated are the FN curve of a fictive plant
and the recommended limit for establishments (F < 10
-3
N
-2
y
-1
for N 10).
Appendix 6.A Probability that the grid point is covered by the cloud, P
ci
In the calculation of the probability of death, P
d
, for toxic clouds we use the probability, P
ci
, that
the grid point is covered by the (effective) cloud. In the first approximation P
ci
is given by:
(6.A.1)
if the grid point is situated in the wind sector and P
ci
(R) = 0 if the grid point is situated outside
the wind sector.
The validity of this approximation is discussed in this appendix.
The origin of this equation can easily be shown in Figure 6.A.1, where the following is valid for
grid point A:
(6.A.2)
The equation is only valid for grid points in the centre of the wind sector and for small cloud
width. However, if the probability of the wind direction, P , does not vary much between
adjacent wind sectors, the equation is also applicable for grid points situated near the boundary
of a wind sector and for large cloud widths.
Figure 6.A.1 Calculation of the probability, P
ci
, that the grid point is covered by the cloud.
If the grid point is situated near the boundary of a wind sector, a combination of two wind
sectors has to be used. For example, in Figure 6.A.1 for grid point B, where < , the
contribution of wind sector II is proportional to:
(6.A.3)
and the contribution of wind sector I is proportional to:
(6.A.4)
If the conditional probability P does not differ much between the two wind sectors, i.e. P
I
P
II
,
the sum of the contributions of the two wind sectors is equal to:
IR
S,M,I
+ IR
S,M,I I
= f
S
P
M
P
cl
(P
I
P
ci,I
+ P
I I
P
ci,I I
)
= f
S
P
M
P
cl
P
I
(P
ci,I
+ P
ci,I I
) (6.A.5)
= f
S
P
M
P
cl
P
I
P
ci
Hence, the approximation is also valid near the boundary of the wind sector, provided that the
conditional probability, P , does not differ much between two wind sectors.
If the effective cloud width, ECW, is larger than the width of the sector, , the
probability that the grid point is covered by the cloud, P
ci
, would be larger than 1. This is, of
course, not a correct value for a probability. However, it can be shown that the formula is still
valid if the conditional probability, P , does not differ much between adjacent wind sectors. In
Figure 6.A.2, a simplified case is shown, in which the contributions of the wind sectors I, II and
III to the risk increment are calculated.
Figure 6.A.2 Calculation of the probability, P
ci
, that the grid point is covered by the cloud for
large values of ECW.
The contribution of the wind sectors I and III to the Individual Risk at grid point A is
proportional to:
, (AI.1)
and the contribution of wind sector II is proportional to:
P
ci,I I
(R) = 1
If the conditional probability P does not differ much between the two wind sectors, i.e. P
I
P
II
P
III
, the sum of the contributions of the three wind sectors is equal to:
IR
S,M,I
+ IR
S,M,I I
+ IR
S,M,I I I
= f
S
P
M
P
cl
(P
I
P
ci,I
+ P
I I
P
ci,I I
+ P
I I I
P
ci,I I I
)
= f
S
P
M
P
cl
P
I
(P
ci,I
+ P
ci,I I
+ P
ci,I I I
)
= f
S
P
M
P
cl
P
I
P
ci
The approximation is also shown to be valid if the effective cloud width, ECW, is larger than the
width of the sector, , provided that the conditional probability, P , does not differ
much between adjacent wind sectors.
It should be noted that the same conclusion holds for the Societal Risk calculation if the
population distribution does not vary much with the wind direction.
If either the probability of the wind direction, P , or the population distribution does vary
considerably between adjacent wind sectors, the contributions of clouds in adjacent wind sectors
have to be taken into account explicitly according to the method outlined here.
Appendix 6.B Sample calculation of the Individual Risk at a grid point
A pipe rupture at the origin results in a continuous release of 100 kg/s CO. The frequency of the
pipe rupture is 5 10
-7
y
-1
. The release takes place at a height of 1 m and z
0
= 0.1 m. Calculate
the contribution of this LOC to the Individual Risk at the grid point (200, 300) using the weather
data of the Rotterdam weather station.
The calculation is carried out for the weather class D 5.0 m s
-1
only.
1. The LOC selected is a pipe rupture leading to a continuous release of 100 kg/s CO. The
frequency of the LOC pipe rupture is f
s
= 5 10
-7
y
-1
.
2. The weather class selected is D 5.0 m s
-1
.
3. The relevant wind sector is the sector 196 - 225 . The probability of weather class D 5.0
m s
-1
and this wind sector occurring is 0.0376 during daytime and 0.0362 during night time
(see Appendix 4.C). Daytime and night-time correspond with fractions 0.44 and 0.56 of
the day, respectively. The probability of the weather class, P
M
P , is therefore
0.44 0.0376 + 0.56 0.0362 = 0.0368.
4. Calculation of the probability of death, P
d
, at the point (100, 200).
4.1 The distance between the grid point (200, 300) and the source at (0,0) is 361 m.
4.2 The concentration on the centre axis is calculated using the basic expressions for the
Gaussian plume model for a continuous release.
(6.B.1)
The calculation is given below, resulting in:
C (x = 361 m, y = 0 m, z = 1 m) = 21.3 g m
-3
4.3 The toxic constants of CO are a = 7.4, b = 1 and n = 1 if the concentration, C, is
given in mg m
-3
and the exposure time, t, in minutes. Using the concentration on the
centre axis, C = 21,300 mg m
-3
, and the maximum exposure time, t = 30 min, the
probit value, Pr = 5.97. The probability of death on the centre line of the plume, P
cl
= 0.835, conforms to Table 5.1.
4.4 The concentration outside the plume axis is given by:
(6.B.2)
The probability of death, P(y), is derived from the concentration, C(x = 361 m, y, z =
1 m), the exposure time, t = 30 min, and the probit function of CO.
5. The PI is given by:
(6.B.3)
Figure 6.B.1 shows the function P(y) and the effective cloud width. In the calculation the
boundaries of the integral are replaced by the distance corresponding to 1% lethality. A
straightforward calculation gives PI = 72 m. The effective cloud width, ECW, is then
calculated as:
Figure 6.B.1 The probability of death, P(y), as a function of the cross-wind distance, y,
and the effective cloud.
5.1 The probability that the grid point (200, 300) is covered by the plume, P
ci
, is given
by:
5.2 The probability of death at the grid point, P
d
, is equal to:
P P P
d cl ci
0381 .
6. The contribution, IR
pipe rupture,M,
, of the LOC pipe rupture, with weather class D 5.0 m
s
-1
and wind direction in sector of 196 - 225 , to the Individual Risk at grid point (200,
300) is therefore:
IR
pipe rupture,M,
= f
pipe rupture
, P
M
P P
d
= 7.0 10
-9
y
-1
7. The total contribution of the LOC pipe rupture to the Individual Risk, IR
pipe rupture
, at
grid point (200, 300) is the sum taken over all (12) wind directions and all (6) weather
classes.
M
(6.B.4)
The contribution of the other wind directions is zero. The contribution of the other weather
classes can be similarly calculated.
Calculation of the concentration for neutral dispersion
The concentration on the centre axis is calculated using the basic expressions for the Gaussian
plume model for a continous release (equations 4.51, 4.53a and 4.57a in [CPR14E]) assuming
total reflection to the ground surface.
The parameters are:
x the co-ordinate along the plume axis, equal to 361 m
y the co-ordinate perpendicular to the plume axis. As the concentration is calculated on the
plume axis, y = 0 m
z the height where the concentration is calculated, equal to 1 m
q the release strength, equal to 100 kg s
-1
h the height of the release, equal to 1 m
u the wind velocity. As the height of the release is less than 10 m, the wind velocity of the
plume is equal to the wind velocity at 10 m, i.e. 5 m s
-1
The dispersion coefficients,
y
en
z
, are calculated using u
*
, yielding the following results:
1/L = 0 m
-1
(Table 4.1, stability: neutral)
u
*
= 0.434 m s
-1
(Equation 4.31, using z = 10 m, z
0
= 0.1 m and u
a
= 5 m s
-1
)
h
i
= 500 m (Table 4.7, using 0.2 x u
*
/ f 770 m)
h = 1 m (the height of the release)
v
= 0.823 m s
-1
(Equation 4.49)
y
(1h) = 41.2 m (Equation 4.54, averaging time 1 h, using t
i
= 300 s, u
a
= 5 m s
-1
and x = 361 m)
y
(10m) = 28.8 m (Equation 4.55, using averaging time 10 min. The value is larger
than the minimum value given in Equation 4.56)
z
= 10.3 m (Equation 4.58a)
Substituting the parameter values results in: C (x = 361 m, y = 0 m, z = 1 m) = 21.3 g m
-3
.
The rule that calculations may terminate at 1% lethality is copied from [KO 20-2].
The presentation of the Individual Risk and Societal Risk is taken from [BRZO]. The minimum
value displayed in the FN-curve used to be 10
-8
y
-1
[BRZO] but has now changed to 10
-9
y
-1
.
Environmental Risk analysis 7.1
7. QUANTITATIVE ENVIRONMENTAL RISK ANALYSIS
A Quantitative Risk Analysis (QRA) is used to demonstrate the risks caused by an activity
involving dangerous substances. The risks calculated in a QRA refer to the external safety, i.e. the
probability of persons off-site dying from exposure. However, an accident with hazardous
substances may also result in damage to the environment. Examples of environmental damage are:
Contamination of groundwater by oil spills. As a result, large areas may become unfit for
drinking-water preparation.
Contamination of surface waters by spills of toxic substances. As a result, life may not be
supported in a watercourse for a prolonged period of time.
Contamination of surface soil by deposition of harmful substances like dioxins and asbestos.
As a result, large areas may become unfit for agriculture and human habitation, and clean-up
may be needed.
A quantitative environmental risk analysis aims at quantifying the risks of environmental impacts
caused by accidental releases.
An accidental release can lead to impacts in three environmental compartments: air, soil (including
groundwater) and surface water. Past accidents with severe environmental consequences have
mostly been related to surface-water pollution. The emphasis of environmental risk
quantification has therefore been in this area. A model to quantify the environmental risks has, to
date, only been developed for the surface water compartment. This model, PROTEUS, is an
aggregation of two models, VERIS and RISAM, which up to now were used in the Netherlands
to evaluate the risks to surface waters.
PROTEUS is a model to calculate the risk to surface waters and sewage treatment plants caused
by an activity with substances which are dangerous to the environment. The risk of an activity
consists of the frequency and the consequences of the spill.
The frequency of a spill is calculated using default failure frequencies, modified by factors for
the technical design of the installation, operational factors (e.g. maintenance of an installation
and working procedures) and management factors. Site-specific information is, in this way,
incorporated into the analysis.
The consequences of a spill depend on the amount of substance spilled and the location-
specific properties of the surface water. The amount of substance spilled depends on the
availability of preventive and mitigating measures.
The results of PROTEUS are presented as a frequencydamage curve similar to the FN-curve in
the QRA. The x-axis represents a measure of environmental damage and the y-axis represents the
cumulative frequency of this environmental damage. Three possible measures of damage can be
chosen:
the amount of substance spilled (in kg).
the area of the surface water affected by the spill (in m
2
or m
3
).
an environmental damage index.
PROTEUS is from end 1998 available as a -version [AVIV98].
Use of new models 8.1
8. THE USE OF NEW MODELS IN A QRA
A Quantitative Risk Analysis (QRA) is used to determine the risk caused by the use, handling,
transport and storage of dangerous substances. The results of the QRA are, for example, used to
assess the acceptability of the risk in relation to the benefits of the activity, to evaluate new
developments on- and off-site, to estimate the benefit of risk-reducing countermeasures and to
determine zoning distances around an activity for land-use planning.
A QRA is intended to give the best estimate of the actual risk level caused by the activity; a QRA
calculation can therefore lead either to an underestimation or an overestimation of the actual risk
level. Since a QRA is intended to give the best estimate, the models used in the QRA represent
the current state of technology and are regularly updated as scientific knowledge increases. New
developments in hardware allow the use of more complex models. The application of an
improved model in a QRA results in either an increase or a decrease in the calculated risk, even if
the actual risk is not changed. This is contrary to a more conservative approach, in which the
calculated risk is assumed to be an overestimation of the actual risk; the application of an
improved model here should result in a reduced overestimation.
The use of continuously improving models in QRAs may lead to problems in the decision-
making process. Examples are:
Zoning distances based on the location of the Individual Risk contours are set for a number of
activities, e.g. transport pipelines and LPG filling stations. The use of an improved model will
result in changes in the location of the Individual Risk contours and the zoning distances may
appear to be no longer correct.
If a new QRA is made to determine the risks caused by an activity, the use of new models can
lead to changes in the calculated risk. Therefore the risk caused by the activity seems to have
changed, although the activity itself and the actual risk have not.
Countermeasures can be taken to reduce the risk. A new QRA is made to determine the effect
of the countermeasures and to quantify the risk reduction achieved. However, the effect of the
countermeasures and the risk reduction achieved can be obscured by changes in the calculated
risk due to the use of new models in the QRA. The result of the new QRA may even show an
increase in the risk calculated despite applying countermeasures.
Recognition of the problems indicated above may lead to the point of view that the models used
in a QRA should be kept fixed to keep in line with previous results and with the decisions based
on these results. However, scientific progress has led to improved estimations of the actual risk.
Consequently, the gap between the best estimate of the risk, as calculated with the newest
models, and the unchanging risk levels, as calculated with the fixed models, increases with time.
Fixing the models can thus lead to other problems:
New models may indicate that larger zoning distances are required for some activities. If the
zoning distances are kept fixed, a situation is created that is considered to be unsafe.
New models may indicate that shorter zoning distances are allowed for some activities. If the
zoning distances are kept fixed, an excessively large area is taken up by the activity no longer
demanded by the risk.
It has therefore been decided to construct a QRA using the current state of technology and the
best models available for this purpose. This is in line with the approach in this guidebook:
although a number of calculation methods are advised, more suitable models can be used when
Use of new models 8.2
available. The adequacy of the QRA and the models used is to be decided by the competent
authorities. Therefore, the user should demonstrate adequate scientific performance in applying
new models to the competent authorities. The scientific performance of the models should be
demonstrated using the results of validation exercises, model intercomparison studies and/or
publications.
If the QRA is made as an update of an existing QRA to incorporate developments on- and off-
site, it is strongly advised to compare the results of the new and existing QRA to facilitate
decision-making processes. The comparison should indicate both the effect of the use of the new
models and of the developments on-site and off-site on the calculated risk.
Uncertainty 9.1
9. UNCERTAINTY IN A QRA
9.1 Introduction
A Quantitative Risk Assessment (QRA) is used to determine the risk caused by an activity
involving dangerous substances. The result of a QRA is the risk around the location of the
activity. Risk is usually presented as a single value, for instance, the frequency of death due to an
accident in the nearby establishment on location (x, y) is 2.3 10
-7
per year. However, QRA
results are calculated with various models having a limited accuracy and for that reason the results
of the QRA calculation have an uncertainty associated with them.
In this chapter, various sources of qualitative uncertainty are discussed, followed by a short
outline on the implications of the evaluation of the QRA results. This chapter is not intended to
give a complete description of all possible sources of uncertainties, but merely to discuss various
types of uncertainty in the QRA. A more elaborate discussion on uncertainties can be found in
the references and references therein, for example [CPR12E, IAEA89, SRPI96].
The term uncertainty is used in this chapter as a measure of distinction between the model
calculation and the actual situation.
9.2 Sources of uncertainty
Various sources of uncertainty are found in a QRA calculation. The sources of uncertainty are
classified here according to the various levels of QRA calculations: starting points, models,
parameter values and the use of the model.
9.2.1 Starting points
Before a QRA calculation is started, choices have to be made with regard to the starting points.
For instance, a conservative approach requires a different type of model and a different set of
parameter values than a best-estimate calculation. Another illustrative example is given as the
following question:
Does the QRA indicate the actual individual risk around an activity, or does the QRA indicate an
artificial individual risk that is independent of the surroundings of the activity?
The consequences of such a choice can easily been seen for releases of flammable substances. If
the QRA is meant to estimate the actual risk, the probability that the flammable cloud will ignite
should be calculated using the location of ignition sources around the activity. On the other hand,
if the QRA is meant to calculate an individual risk independent of the surroundings of the
activity, the ignition sources around the activity should be ignored and agreements will have to be
made on the ignition of flammable clouds.
This guideline has established a number of starting points and in doing so minimises the
associated uncertainties. However, differences in starting points may hamper the comparison of
new QRAs with those made previously and the comparison of QRAs made for establishments
with those made for transport activities.
Uncertainty 9.2
9.2.2 Models
When the starting points of the QRA are defined, appropriate models have to be used.
Uncertainties arise from various sources:
A number of potentially important processes may be ignored in the models. Examples are
tank-roof collapse, atmospheric deposition processes and chemical reactions in the dispersing
cloud.
The models are not valid for the specific local situation. For instance, the dispersion models
used are only valid for a flat terrain in the absence of large obstacles, whereas numerous
obstacles can be present in and around industrial sites.
Processes are often simplified in models. For instance, sometimes a uniform wind speed is
used, ignoring the variation in wind speed with height.
Natural variability is ignored. As an example, all humans are assumed to react similarly to an
exposure to toxic substances. However, elderly people are probably more vulnerable and thus
are at greater risk.
Models are sometimes used outside the range of applicability. For instance, dispersion
models that are validated for neutral gas dispersion are sometimes used to calculate the
dispersion of heavy or light gases. Dispersion models are occasionally also used at distances
outside the validity range of the model.
The computer code contains numerical approximations. If the numerical step in space and
time is too large, errors are introduced. Also coding errors can be present in the source code of
the computer program.
It should be noted that the models for use in the QRA are not strictly established. Different
models, varying in complexity and accuracy, can be used in a QRA calculation,. The results of the
QRA can depend strongly on the models used (see, for example, [TNO98c]). As a result, the
selection of the models in the QRA can determine the assessment of the acceptability of the
activity, a problem already touched on in Chapter 8 describing the use of new developments in
scientific knowledge.
9.2.3 Parameter values
Before the model can be used, input parameter values are collected from either the literature or
experiments. Sources of uncertainty in parameter values are:
Parameter values are extrapolations of measured data. Toxicity parameters are often derived
from animal toxicity data that are measured at a different dose level. Extrapolation occurs from
animal to human toxicity and from the doses applied in the experiment to the doses relevant
for QRA calculations. Physical data, like the flash fraction, are derived from small-scale
experiments and need to be extrapolated to large-scale installations.
Generic data sets are used in the absence of location specific data. For example, often generic
data are used for parameters like the roughness of a pipe wall and the failure frequency of a
pipeline, since specific data are either not available or too difficult to obtain.
The data to derive parameter values may be sparse. For instance, failure frequencies of storage
tanks are derived from historical data sets. Since accidents are rare, failure frequencies of
installations have large uncertainties. Furthermore, databases describe the historical failure rate
and these data are possibly no longer valid for present-day installations.
Uncertainty 9.3
9.2.4 Use of the model
Finally, the models are all used in the QRA. However, different users may have different results
using the same model for the same installation, for example:
The user may misinterpret the model input and output due to inexperience.
The user can make calculation errors, errors in manual copying of results and in the reporting
of the results.
The user may reduce the calculation time by reducing the number of calculations, e.g. by
selecting only a limited number of installations.
9.3 Quantification of uncertainties
To determine the magnitude of some sources of uncertainty and the consequences to the results
of the QRA, various tools are available. For example, statistical analysis of historical data gives
information on the uncertainty in the failure frequencies derived. The consequences of parameter
uncertainty to the model outcome can be determined with the use of sensitivity and uncertainty
analysis.
The uncertainty in models can partly be established by model validation studies. In these studies,
model predictions are compared with experimental data sets. Alternatively, QRA-type models
can be compared to more complex models in model intercomparison studies.
Uncertainties in the results of a QRA are not considered in the decision-making process since the
criteria for the evaluation of the acceptability of risk are expressed as single numbers, e.g. at
location (x,y) the frequency of dying due to an accident must be less than 1 10
-6
yr
-1
. The
results of a QRA calculation are therefore also single numbers, which may either overestimate or
underestimate the actual risk level due to uncertainties. However, the risk calculated should be a
best estimate of the actual risk level and the reliability of the risk calculation should be secured as
much as possible by the use of accurate calculation methods.
References 10.1
10. REFERENCES
[AEC75] US Atomic Energy Commission. Reactor safety study - an assessment of accident
risks in the U.S. Commercial Nuclear Power Plants; WASH-1400, Appendix III
Failure data (and references therein), 1975.
[AM94] AMINAL Dienst Gevaarlijke Stoffen en Risicobeheer. Handboek kanscijfers ten
behoeve van het opstellen van een veiligheidsrapport. Brussels: AMINAL, 1994.
[AVIV98] AVIV, RIZA, RIVM, VROM. PROTEUS V1.00. Enschede: AVIV, 1998.
[BR97] Bridgis. Population data for the Netherlands in the year 1996. Druten: Bridgis, 1997.
[BRZO] Besluit risicos zware ongevallen (besluit van 15 september 1988, Stb. 432; laatstelijk
gewijzigd bij besluit van 20 juni 1994, Stb. 463). The Hague: SDU, 1988/1994.
[Bu75] Bush S.H. Pressure Vessel Reliability. Trans. of AMSE - Journal of Pressure Vessel
Technology, 1975.
[Ch88] Chambers science and technology dictionary. Edingburgh: W & R Chambers Ltd and
Cambridge University Press, 1988
[COVO81] COVO Commission. Risk analysis of six potentially hazardous industrial objects in
the Rijnmond area, a pilot study. A report to the Rijnmond public authority.
Schiedam: Central Environmental Control Agency Rijnmond, 1981.
[CPR12E] Committee for the Prevention of Disasters. Methods for determining and processing
probabilities (Red Book), Second edition. The Hague: SDU, 1997.
[CPR14] Committee for the Prevention of Disasters. Methoden voor het berekenen van
fysische effecten. Voorburg: Ministry of Social Affairs and Employment, 1988
2
.
[CPR14E] Committee for the Prevention of Disasters. Methods for the calculation of physical
effects (the Yellow Book). The Hague: SDU, 1997
3
.
[CPR15] Ministry of Housing, Spatial Planning and the Environment. Risico-analyse
methodiek CPR-15 bedrijven. The Hague: Ministry of Housing, Spatial Planning and
the Environment, 1997.
[CPR16] Committee for the Prevention of Disasters. Methods for the calculation of damage
(the Green Book). Voorburg: Ministry of Social Affairs and Employment, 1990
1
[DIPPR] Daubert T.E. and Danner R.P. (eds.). DIPPR. Physical and Thermodynamic
Properties of Pure Chemicals: Data Compilation. New York: Hemisphere Publishing
Corp., 1992.
[DNV96] DNV Technica. SAFETI 3.4 Theory Manual. London: DNV Technica, 1996.
References 10.2
[DNV98] DNV Technica. PHAST 5.2. London: DNV Technica, 1998.
[EU88] Richtlijn van de Raad van 7 juni 1988 betreffende de onderlinge aanpassing van de
wettelijke en bestuursrechterlijke bepalingen van de Lid-Staten inzake de indeling, de
verpakking en het kenmerken van gevaarlijke preparaten (88/379/EEG). Luxembourg:
Office for official publications of the EC, 1988.
[EU96] Council Directive 96/82/EC of 9 December 1996 on the control of major-accident
hazards involving dangerous substances. Luxembourg: Office for official publications
of the EC, 1996.
[EU98] Commission Decision of ___ 1998 on harmonized criteria for dispensations
according to article 9 of Council Directive 96/82/EC of 9 December 1996 on the
control of major-accident hazards involving dangerous substances. Luxembourg: Draft
20.3.1998.
[Gi97] Gilham S., Deaves D.M., Hoxey R.P., Boon C.R. and Mercer A. Gas build-up within
a single building volume - comparison of measurements with both CFD and simple
zone modelling. Journal of Hazardous Materials 53: 93-114, 1997.
[Hu92] Hurst N.W., Hankin R.K.S., Wilkinson J.A., Nussey C., Williams J.C. Failure rate
and incident databases for major hazards. Published in the Proceedings of the 7th
International Symposium on Loss Prevention and Safety Promotion in the Process
Industry, Taormina, Italy 4 - 8 may 1992. Rome: SRP Partners, 1992.
[IAEA89] International Atomic Energy Agency. Evaluating the reliability of predictions made
using environmental transfer models. Vienna: IAEA, 1989.
[IPO] IPO A73. Handleiding voor het opstellen en beoordelen van een extern
veiligheidsrapport. IPO Project A73. The Hague: IPO, 1994.
Committee for the Prevention of Disasters, Risk Evaluation. Publication RE-95-1;
Tekstwijzigingen op de IPO A73. The Hague: Ministry of Housing, Spatial Planning
and the Environment, 1996.
[IPORBM]AVIV. IPO RisicoBerekeningsMethodiek. Enschede: AVIV, 1997.
[Ja71] Jacobs R.M. Minimizing hazards in design. Quality Progress, 1971.
[KNMI72] KNMI (Royal Dutch Meteorological Institute). Climatological data of Netherlands
stations. No. 8. Frequency tables of atmospheric stability. De Bilt: KNMI, 1972.
[KNMI92] KNMI (Royal Dutch Meteorological Institute). Klimatologische gegevens van
Nederlandse stations ; normalen en extreme waarden van 15 hoofdstations voor het
tijdvak 1961-1990. De Bilt: KNMI, 1992.
[KO 9] Knelpuntenoverleg EVR. Aanwijzingssysteem per inrichtingsonderdeel. KO 9. The
Hague: Ministry of Housing, Spatial Planning and the Environment , 1989.
References 10.3
[KO 12] Knelpuntenoverleg EVR. Laad en loshandelingen. KO 12. The Hague: Ministry of
Housing, Spatial Planning and the Environment , 1989.
Knelpuntenoverleg EVR. Laad en loshandelingen. KO 20-2. The Hague: Ministry of
Housing, Spatial Planning and the Environment , 1989.
[KO 19-2] Knelpuntenoverleg EVR. Toepassen van het subselectie systeem bij een EVR-
plichtige inrichting met schepen. KO 19-2. The Hague: Ministry of Housing, Spatial
Planning and the Environment, 1989.
[KO 20-2] Het gebruik van parameters in de risico-analyse. KO 20-2. The Hague: Ministry of
Housing, Spatial Planning and the Environment, 1989.
[KO 22-5] Knelpuntenoverleg EVR. QRA ladende en lossende schepen. KO 22-5. The Hague:
VROM, 1990
[KO 24-2] Gebruik toxiciteitsgegevens. KO 24-2. The Hague: Ministry of Housing, Spatial
Planning and the Environment 1989.
[KO 59] Gebruik toxiciteitsgegevens. Ammoniakprobitfunktie. KO 59. The Hague: Ministry
of Housing, Spatial Planning and the Environment, 1992.
[KO 86] Gebruik toxiciteitsgegevens. Probitfunctie fosgeen. KO 56. The Hague: Ministry of
Housing, Spatial Planning and the Environment, 1994.
[NATO92] Manual of NATO safety principles for the storage of military ammunition and
explosives (AC258), Allied Ammunition Storage and Transport Publication 1
(AASTP-1), May 1992.
[Le95] Leeuwen C.J. van, and Hermens J.L.M. (eds.). Risk assessment of chemicals: an
introduction. Dordrecht: Kluwer Academic Publishers, 1995.
[NM86] Werkgroep Verspreiding Luchtverontreiniging. Invloed van een gebouw op de
verspreiding van schoorsteenpluimen. Delft: Vereniging Lucht, 1986.
[NNM98] Het Nieuw Nationaal Model. Infomil: Den Haag, 1998.
[NR] Nadere Regels met betrekking tot rapport inzake de externe veiligheid (ministerile
regeling d.d. 3 februari 1989, Stcrt. 31). The Hague: SDU, 1989.
[P 172] Arbeidsinspectie. Arbeidsveiligheidsrapport; leidraad aanwijzing AVR-plichtige
installaties. Publikatie P 172-1. Voorburg: Ministry of Social Affairs and
Employment, 1988
1
.
[Pe84] Perry R.H., Green D.W. and Maloney J.O. (eds.). Perrys Chemical engineers
handbook. New York: McGraw-Hill, 1984
6
.
References 10.4
[Ph69] Phillips C.A.G. and Warwick R.G. A survey of defects in pressure vessels built to
high standards of construction and its relevance to nuclear primary circuits. UKAEA
AHSB(S) R162, 1969.
[Re88] Reid R.C., Prausnitz J.M. and Poling B.E. The properties of gases and liquids. New
York: McGraw-Hill, 1988
4
.
[RIVM99] RIVM. SERIDA. Safety Environmental Risk Database. Bilthoven: RIVM, 1999.
[SA75] Science Applications Inc. LNG terminal risk assessment study for Los Angeles,
California. Report prepared for western LNG Terminal Company, 1975
[SAVE97] SAVE. Handleiding protocol voor het uitvoeren van een QRA voor
goederenemplacementen. Apeldoorn: SAVE , 1997
[Sm74] Smith T.A. and Warwick R.G. The second survey of defects in pressure vessels built
to high standards of construction and its relevance to nuclear primary circuits. Safety
and Reliability Directorate, SRD R30, 1974.
[SRPI96] BIOMOVS II Steering Committee. An overview of the BIOMOVS II Study and its
Findings. Stockholm; Swedish Radiation Protection Institute, 1996.
[SRS] SRS Data bank.
[SZW97] Beleidsregels arbeidsomstandighedenwetgeving. Besluit van de Staatssecretaris van
Sociale Zaken en Werkgelegenheid van 27 juni 1997, Directie Arbeidsomstandigheden,
Arbo/AIS 9701436 tot vaststelling van beleidsregels op het gebied van de
Arbeidsomstandighedenwetgeving (Beleidsregels arbeidsomstandighedenwetgeving).
Supplement van de Nederlandse Staatscourant van 27 juni 1997. The Hague: SDU,
1997.
[Ta98] Taylor J.R./ Taylor Associates ApS. Review of failure rate data for risk analyses.
Version 1 Issue 1. Glumsoe: Neste and Taylor Associates ApS, 1998.
[Ti97] Timmers, P.G.J. Berekening van het in- en extern risico van explosievenopslag met
behulp van RISKANAL. Draft, Rijswijk: TNO, 1997.
[TNO83] TNO. LPG Integraal, rapport 1112 Effectmodellen LPG. Apeldoorn: TNO, 1983.
[TNO98a] TNO. Ventilatieonderzoek naar infiltratie en verspreiding van
buitenluchtverontreinigingen in woningen bij calamiteiten. Delft: TNO, 1998.
[TNO98b] Logtenberg, M.Th.. Derivation of failure frequencies for LoC cases. TNO report
TNO-MEP-R98/501. Apeldoorn: TNO, 1998.
[TNO98c] TNO. Investigation into the consequences of introduction of the third version of the
yellow book. TNO Report TNO-MEP-R98/457. Apeldoorn: TNO, 1998.
References 10.5
[TNO98d] J.M. Ham and H.H. Schoten. Damage determining effects of BLEVEs. TNO-MEP
note 1999IV/38-29171/HAJ/mba. Apeldoorn: TNO, 1998.
[VITO97] Gheys K.. Onderzoek van berekeningsmethoden voor fysische effecten van het
incidenteel vrijkomen van gevaarlijke stoffen. Eindrapport. VITO: Antwerpen, 1997.
[VVoW95] Systematiek voor indeling van stoffen ten behoeve van risico-berekeningen bij het
vervoer van gevaarlijke stoffen. Project Veilig Vervoer over Water; Deelproject S3b.
The Hague: Ministry of Transport, Public Works and Water Management, 1995
[We76] Welker J.R. et al. Fire safety aboard LNG vessels. NTIS AD/A - 030 619, 1976
[Wi98] Witlox H.W.M. Modelling of building-wake dispersion. Londen: DNV Technica,
1998.
[Ya77] Yaws C.L. Physical Properties. New York: McGraw-Hill, 1977.
Glossary G.1
GLOSSARY
1% lethality the boundary where 1% of the population exposed suffers
death due to an accident with dangerous substances
atmospheric tank storage vessel in which the pressure is close to one bar absolute
blanketing applying a layer of inert gas over a dangerous substance
blast (wave) a rapidly propagating pressure or shock-wave in the
atmosphere, with high pressure, high density and high particle
velocity
blast strength measure used in the multi-energy method to indicate the
strength of the blast, indicated by a number ranging from 1 (for
very low strengths) up to 10 (for detonative strength)
BLEVE Boiling Liquid Expanding Vapour Explosion; results from the
sudden failure of a vessel containing liquid at a temperature well
above its normal (atmospheric) boiling point. A BLEVE of
flammables results in a large fire ball.
blocking system repression system to isolate (part of) an installation to prevent
outflow
bund containment area to restrict spread of liquid
carcinogenic carcinogenicity; capacity of a chemical to induce cancer
CFD calculation three dimensional model calculation; the dispersion of gas is
calculated by solving the dynamic partial differential equations
for a set of control volumes
competent authority authority licensing the activity with dangerous substances
dense gas gas which has a higher specific weight than the surrounding
ambient air
deposition absorption of gas or particles by the ground or vegetation
design pressure pressure for which the installation is designed; the installation
should be able to withstand this pressure
detonation a propagating chemical reaction of a substance in which the
propagation of the reaction front is determined by compression
beyond the auto-ignition temperature
dispersion mixing and spreading of gases in air, causing clouds to grow
Glossary G.2
domino effect the effect that loss of containment of one installation leads to
loss of containment of other installations
dose a measure of integral exposure
dry deposition deposition not caused by rain
effective cloud width the width of a uniform cloud replacing a toxic cloud; the uniform
cloud has a constant probability of death equal to the
probability of death on the centre line of the toxic cloud and the
same probability integral
entrainment mixing of (clean) air into a cloud or plume
establishment the whole area under the control of an operator where dangerous
substances are present in one or more installations, including
common or related infrastructures or activities
event tree a logic diagram of success and failure combinations of events
used to identify accident sequences leading to all possible
consequences of a given initiating event
explosion a sudden release of energy that causes a blast
exposure concentration or intensity that reaches the target person, usually
expressed in terms of concentration or intensity and duration
failure a system or component failure occurs when the delivered
service deviates from the intended service
fault tree analysis the evaluation of an undesired event, called the top event of the
fault tree. Given the top event, a fault tree is constructed by a
deductive (top-down) method of analysis, identifying the cause
or combination of causes that can lead to the defined top event
fire ball a fire, burning rapidly enough for the burning mass to rise into
the air as a cloud or ball
flash evaporation see flashing
flashing part of a superheated liquid that evaporates rapidly due to a
relatively rapid depressurisation, until the resulting
vapour/liquid mixture has cooled to below boiling point at the
end pressure. Superheat is the extra heat of a liquid made
available by decreasing the liquids temperature, for instance, by
vaporisation, until the vapour pressure equals that of the
surroundings.
Glossary G.3
flash fire the combustion of a flammable vapour and air mixture in which
the flame passes through the mixture at a rate less than sonic
velocity so that negligible damaging overpressure is generated
FN-curve log-log graph, where the x-axis represents the number of deaths,
N, and the y-axis represents the cumulative frequency of the
accidents, with the number of deaths equal to N or more
free field calculation calculation method in which ignition sources outside the
establishment or transport route are not taken into account. If a
flammable cloud is not ignited at the establishment, ignition is
assumed to occur at maximum cloud area.
frequency the number of times an outcome is expected to occur in a given
period of time (see also probability)
friction velocity by definition the cube root from (minus) the shear stress at the
surface; the shear stress is the stress exerted by the wind on the
ground surface due to friction
grid A network of lines superimposed on a map and forming squares
for referencing, the basis of the network being that each line in it
is found at a known distance either east or north of a selected
origin
grid cell the area around a single grid point corresponding with the mesh
of the grid
grid point crossing point of two grid lines
hazard a chemical or physical condition with the potential of causing
damage
ignition source a thing able to ignite a flammable cloud, e.g. due to the presence
of sparks, hot surfaces or open flames
indication number measure of the hazard of an installation, irrespective of its
location
individual risk the probability that in one year a person will become a victim o
an accident if the person remains permanently and unprotected
in a certain location. Often (also in this report) the probability
of occurrence in one year is replaced by the frequency of
occurrence per year.
installation a technical unit within an establishment in which dangerous
substances are produced, used, handled or stored.
Glossary G.4
internal energy energy present due to storage of a substance above ambient
pressure; following rupture of the vessel, part of the internal
energy is released as a blast wave
jet the outflow of material emerging from an orifice with a
significant momentum
jet fire see jet flame
jet flame the combustion of material emerging from an orifice with a
significant momentum
K1 liquid flammable liquid having a flash point less than 21 C and a
vapour pressure at 50 C less than 1.35 bar (pure substances)
or 1.5 bar (mixtures)
LC
50
median lethal concentration , i.e. the concentration of a
substance estimated to be lethal to 50% of the test organisms.
LC
50
(rat, inh, 1 h) is the concentration in air estimated to be
lethal to rats after one hour of exposure.
lee location/site downwind of a building
LFL lower flammability limit; below this concentration too little
flammable gas is present in the air to maintain combustion
limit value measure of the dangerous properties of a substance based on
both the physical and the toxic/explosive/flammable properties
of the substance
liquid head vertical distance between the liquid level and the location of the
hole
liquefied pressurized gas see pressurized liquefied gas
Loss of Containment event event resulting in the release of material to the atmosphere
LOC see Loss of Containment event
maximum occurring pressure maximum pressure that can occur in an installation
model intercomparison comparison of the results of two models for a test case
Monin-Obukhov length length-scale which characterises the atmospheric stability
nominal pumping rate normal flow of material through a pump
non-obstructed region region where jet outflow does not interact with objects
Glossary G.5
obstructed outflow region where a jet outflow does interact with objects
operator any individual or corporate body who operates or holds an
establishment or installation or, if provided for by national
legislation, has been given decisive economic power in the
technical operation thereof (Chapter 2); also defined as any
individual operating technical equipment (Chapter 4)
outflow see release
Pasquill class classification to qualify the stability of the atmosphere,
indicated by a letter ranging from A, for very unstable, to F, for
stable
passive dispersion dispersion solely caused by atmospheric turbulence
plume cloud of material following a continuous release to the
atmosphere
pool layer of liquid on a subsoil or on a water surface
pool fire the combustion of material evaporating from a layer of liquid at
the seat of the fire
pressure vessel storage vessel in which the pressure is (substantially) more than
1 bar absolute
pressure relief device valve or bursting disc, designed to relieve excessive pressure
automatically
pressurized liquefied gas gas that has been compressed to a pressure equal to saturated
vapour pressure at storage temperature, so that the larger part
has condensed to the liquid state
primary container container holding the substance and in direct contact with it
probability measure of the likelihood of an occurrence, expressed as a
dimensionless number between 0 and 1. Risk is defined as the
probability that within a fixed time period, usually one year, an
unwanted effect occurs. Consequently, risk is a dimensionless
number. However, risk is often expressed in units of frequency,
per year. Since failure frequencies are low, the probability that
an unwanted effect will occur within a fixed time period of one
year is, practically speaking, equal to the frequency of
occurrence per year. In this report, frequency is used to denote
the risk.
probability integral integral of the probability of death over the co-ordinate
perpendicular to the plume axis
Glossary G.6
probit number directly related to probability by a numerical
transformation
process vessel vessel in which a change in the physical properties of the
substance occurs, e.g. temperature or phase
PROTEUS model to calculate the risk to surface waters and sewage
treatment plants
puff cloud spreading in all directions due to an instantaneous release
QRA see Quantitative Risk Assessment
quantitative risk assessment the process of hazard identification followed by a numerical
evaluation of effects of incidents, and consequences and
probabilities, and their combination into overall measures of ris
k
rain-out dropping to the ground of the small liquid drops from that
fraction of the flashing liquid initially remaining suspended in
the atmosphere
reactive measure for the flame acceleration in a gasair mixture
reactor vessel vessel in which a chemical change of the substances occurs
release the discharge of a chemical from its containment, i.e. the proces
s
and storage equipment in which it is kept
recirculation zone location/site downwind of a building where uniform mixing is
assumed
repression system system to limit the release of substances into the environment
given a loss of containment event
RISAM model to calculate the risk to surface waters
risk the unwanted consequences of an activity connected with the
probability of occurrence. Often (also in this report) the
probability of occurrence is replaced by the frequency of
occurrence
risk contour line on a map connecting points having equal risk
roughness length artificial length-scale appearing in relationships describing the
wind speed over a surface and characterising the roughness of
the surface
Glossary G.7
safety report report on the safety of an establishment, as required by Council
Directive 96/82/EC of 9 December 1996
saturation pressure the pressure of a vapour which is in equilibrium with its liquid
state; also the maximum pressure possible for a vapour at a
given temperature
secondary container container enclosing the primary container, not in direct contact
with the substance
selection number measure of the hazard of an installation at a specific location
side-on overpressure the pressure experienced by an object as a blast-wave passes by
societal risk the frequency (per year) that a group of at least a certain size
will at one time become victims of an accident
stability Atmospheric stability; the extent to which vertical temperature
(= density) gradients promote or suppress turbulence in the
atmosphere
surface friction velocity see friction velocity
sympathic detonation domino-effect, where detonation of explosives in one storage
room leads to detonation of explosives in other storage rooms
uncertainty measure of the distinction between the model calculation and
the actual situation
validation Comparison of model results to measurements
vapour cloud explosion the explosion resulting from ignition of a pre-mixed cloud of
flammable vapour, gas or spray with air, in which flames
accelerate to sufficiently high velocities to produce significant
overpressure
VERIS model to calculate the risk to surface waters
wall roughness measure (in metres) to express the influence of the pipe wall on
the flow through the pipe
wet deposition deposition caused by rain
Note: Several definitions have been taken from the Red Book [CPR12E], the Yellow Book
[CPR14E], Chambers [Ch88], Council Directive 96/82/EC [EU96] and Van Leeuwen
and Hermens [Le95].
Symbols S.1
SYMBOLS
Chapter 2
Amount added in the calculation of the factor O
3
to account for liquid pool
evaporation (-)
A Indication number (-)
A
F
Indication number for flammable substances (-)
A
T
Indication number for toxic substances (-)
G Limit value (kg)
L Distance between an installation and the location where a selection number is
calculated (m)
LC
50
Lethal concentration at which 50% of the exposed test animals suffer death (mg m
-3
)
I
i
Sample installation (-)
O
1
Factor to account for the type of installation, storage or process (-)
O
2
Factor to account for the positioning of the installation (-)
O
3
Factors to account for the process conditions (-)
P
i
Partial vapour pressure at process temperature (bar absolute)
P
sat
Saturation pressure at process temperature (bar absolute)
Q Quantity of a substance present (kg)
Q
x
Qualifying quantity of substance x for application of Article 9 of Council Directive
96/82/EC of 9 December 1996 (tonnes)
q
x
Quantity of substance x present or likely to be present for application of Article 9 of
Council Directive 96/82/EC of 9 December 1996 (tonnes)
S Selection number (-)
S
T
Selection number for toxic substances (-)
S
F
Selection number for flammable substances (-)
T
p
Process temperature ( C)
T
bp
Atmospheric boiling point ( C)
X Factor in O
3
to account for saturation pressure at process temperature (-)
Symbols S.2
Chapter 4
Note: The symbols used in Appendix 4.A are listed there.
Adiabatic flash fraction (-)
Wall roughness (m)
TNT-equivalence factor (-)
s
Thermal conductivity (J s
-1
m
-1
K
-1
)
s
Density (kg m
-3
)
Ignition effectiveness (s
-1
)
A Projection of a building in wind direction (m
2
)
A
bund
Bund area (m
2
)
a
s
Thermal diffusivity (m
2
s
-1
)
b
0
Source radius of the free jet (m)
C
d
Discharge coefficient (-)
C
rz
Concentration in the recirculation zone (kg m
-3
)
C
u
Empirical constant used to calculate the length of the free jet (-)
C
vent
Concentration in ventilation air (kg m
-3
)
c
p,s
Specific heat (J kg
-1
K
-1
)
D
rel,i
Release duration of time segment i (s)
d Average traffic density (-)
E Combustion energy (J)
F Ventilation rate (m
3
s
-1
)
f Frequency (y
-1
)
f
obstr
Fraction of the total mass in the flammable cloud in an obstructed region (-)
K Parameter to account for the shape and orientation of the building (-)
L Monin Obukhov length (m)
L
0
Distance between release point and obstacle (m)
L
j
Length of free jet (m)
M Mass released (kg)
M
rel
Mass released in the first 30 minutes following the LOC (kg)
M
seg
Mass released in a time segment (kg)
N Number of vehicles per hour (h
-1
)
N
seg
Number of time segments (-)
n Average number of people present in a grid cell (-)
Symbols S.3
P(t) Probability of ignition in time interval 0 t (-)
P
a
Ambient pressure (N m
-2
)
P
BLEVE
Probability of a BLEVE given an instantaneous release of a flammable gas with direct
ignition (-)
P
d.i.
Probability of direct ignition (-)
P
i
Probability of an obstructed outflow (-)
P
present
Probability that an ignition source is present (-)
Q
out
Mass outflow to the atmosphere following a release inside a building (kg s
-1
)
Q
rel,i
Mass outflow in time segment i (kg s
-1
)
R
0.3 barg
Distance to the peak- side on overpressure contour of 0.3 barg (m)
R
0.1 barg
Distance to the peak- side on overpressure contour of 0.1 barg (m)
R
pool
Effective pool radius (m)
r
0.3 barg
Combustion energy scaled distance to the peak- side on overpressure contour of 0.3
barg (-)
r
0.1 barg
Combustion energy scaled distance to the peak- side on overpressure contour of 0.1
barg (-)
s Co-ordinate along the jet axis (m)
t Time (s)
t
av
Averaging time (s)
t
react
Time for a repression system to become effective (s)
t
rel
Duration of a release (s)
u Wind speed (m s
-1
)
u
*
Friction velocity (m s
-1
)
u
0
Velocity of the jet at the source (m s
-1
)
u
air
Average ambient wind velocity (m s
-1
)
u
c
Velocity of the jet (m s
-1
)
V Volume of a room (m
3
)
v Average velocity of vehicle (km h
-1
)
z
0
Roughness length (m)
Symbols S.4
Chapter 5
a Probit constant describing the toxicity of a substance (-)
b Probit constant describing the toxicity of a substance (-)
C Concentration of a toxic substance (mg m
-3
)
F
E
Fraction of the population dying (-)
F
E,in
Fraction of the population indoors who die (-)
F
E,out
Fraction of the population outdoors who die (-)
f
pop,in
Fraction of population present indoors (-)
f
pop,out
Fraction of population present outdoors (-)
n Probit constant describing the toxicity of a substance (-)
Q Heat radiation (W m
-2
)
P Probability (-)
P
E
Probability of death (-)
P
i
Probability of ignition event i (-)
P
peak
Peak overpressure (bar gauge)
Pr Probit (-)
t Exposure time (min)
Chapter 6
IR
S,M, ,i
Contribution to the individual risk of LOC, S, weather class, M, wind direction, ,
and ignition event, i (y
-1
)
N
S,M, ,i
Expected number of deaths in a grid cell of LOC, S, weather class, M, wind
direction, , and ignition event, i (-)
T Time step (s)
angle (-)
angle (-)
angle (-)
y
,
z
Dispersion coefficients (m)
C(R,t) Concentration of a toxic substance on the centre line (mg m
-3
)
ECW(R) Effective cloud width at distance R (m)
F
cl
(R) Fraction of deaths on the centre line at distance R (-)
F
d
Fraction of people dying, given LOC, S, weather class, M, wind direction, , and
ignition event, i (-)
Symbols S.5
F
E,in
Fraction of the population indoors who die(-)
F
E,out
Fraction of the population outdoors who die (-)
F
N
Summed frequency of events with N or more people dying (y
-1
)
f
pop,in
Fraction of population present indoors (-)
f
pop,out
Fraction of population present outdoors (-)
f
S
Frequency of LOC, S (y
-1
)
f
S,M, ,i
Frequency of LOC, S, weather class, M, wind direction, , and ignition event, i (y
-1
IR Individual risk at a point (y
-1
)
N
cell
Number of people in a grid cell (-)
N
S,M, ,i
Expected number of deaths of LOC, S, weather class, M, wind direction, , and
ignition event, i (-)
n
ws
Number of wind sectors (-)
P(R,y) Probability of death off-axis at distance R (-)
P Probability of wind direction (-)
P
ci
Probability a grid point is covered by the effective cloud (-)
P
cl
(R) Probability of death on the centre line at distance R (-)
P
d
Probability of death at a point (-)
P
M
Probability of weather class M (-)
PI(R) Probability integral at distance R (m)
R Distance between grid point and source (m)
x Co-ordinate along the centre line of the plume (m)
y Co-ordinate perpendicular to the centre line of the plume (m)
z Height co-ordinate (m)
Guideline
Ior
Quantitative Risk Assessment
Part two: Transport
ir. G.W.M. Tiemessen
ir. J. Heitink
AVIV
ir. B.L. Wiekema
dr.ir. L.A.M. Janssen
SAVE
TABLE OF CONTENTS
MANUAL FOR RISK ANALYSIS OF TRANSPORTATION ACTIVITIES, PREFACE
1. SELECTION OF RELEVANT ROUTE SECTIONS
1.1 INTRODUCTION ................................................................................................................ 1.1
1.2 THRESHOLD VALUES EOR ROADS ....................................................................................... 1.1
1.2.1 Individual risk ....................................................................................................... 1.2
1.2.2 Societal risk ........................................................................................................... 1.2
1.3 THRESHOLD VALUES EOR RAILWAYS .................................................................................. 1.4
1.3.1 Individual risk ....................................................................................................... 1.4
1.3.2 Societal risk ........................................................................................................... 1.5
1.4 THRESHOLD VALUES EOR INLAND WATERWAYS ................................................................... 1.6
1.4.1 Individual risk ....................................................................................................... 1.7
1.4.2 Societal risk ........................................................................................................... 1.8
2. IPORBM
2.1 INTRODUCTION ................................................................................................................ 2.1
2.2 IPORBM: A GENERAL DESCRIPTION ................................................................................ 2.1
2.3 ROAD .............................................................................................................................. 2.3
2.4 RAIL ............................................................................................................................... 2.4
2.5 INLAND WATERWAY .......................................................................................................... 2.8
2.6 PIPELINE ........................................................................................................................ 2.10
3. DETAILED QRA
3.1 INTRODUCTION ................................................................................................................ 3.1
3.2 ROAD TRANSPORT ............................................................................................................ 3.2
3.2.1 Loss oI containment events ................................................................................... 3.2
3.2.2 Accident and outIlow Irequency, events and event probabilities .......................... 3.2
3.2.3 Modelling source term, dispersion, exposure and damage ................................... 3.5
3.2.4 Calculation and presentation oI results ................................................................. 3.7
3.2.5 Obstacles ............................................................................................................... 3.8
3.3 RAIL TRANSPORT ............................................................................................................. 3.9
3.3.1 Introduction ........................................................................................................... 3.9
3.3.2 Loss oI containment events ................................................................................... 3.9
3.3.3 Accident and outIlow Irequency, events and event probabilities ........................ 3.10
3.3.4 Modelling source term, dispersion, exposure and damage ................................. 3.14
3.3.5 Calculation and presentation oI results ............................................................... 3.15
3.4 INLAND WATERWAY TRANSPORT ...................................................................................... 3.15
3.4.1 Introduction ......................................................................................................... 3.15
3.4.2 Loss oI containment events ................................................................................. 3.16
3.4.3 Accident Irequency and outIlow probabilities .................................................... 3.18
3.4.6 Presentation oI results ......................................................................................... 3.23
3.5 PIPELINE TRANSPORT ..................................................................................................... 3.23
3.5.1 Introduction ......................................................................................................... 3.23
Table oI contents
3.5.2 Loss oI Containment Events ............................................................................... 3.24
3.5.3 Accident and outIlow Irequency, events and event probabilities ........................ 3.24
4. BASIC DATA
4.1 INTRODUCTION ................................................................................................................ 4.1
4.2 CATEGORISING SUBSTANCES ............................................................................................. 4.1
4.2.1 Description oI method and criteria ........................................................................ 4.1
4.2.2 Use oI method in a QRA ....................................................................................... 4.3
4.3 ROAD TRANSPORT ............................................................................................................ 4.4
4.3.1 Transported substances .......................................................................................... 4.4
4.3.2 Determination oI location speciIic outIlow Irequencies ....................................... 4.5
4.4 RAILWAY TRANSPORT ....................................................................................................... 4.6
4.4.2 Determination oI location speciIic outIlow Irequencies ....................................... 4.6
4.5 INLAND WATERWAY TRANSPORT ........................................................................................ 4.6
4.5.2 Determination oI location speciIic accident Irequencies ...................................... 4.8
4.6 PIPELINE TRANSPORT ..................................................................................................... 4.10
4.6.1 Transported substances......................................................................................... 4.10
4.6.2 Determination oI location speciIic outIlow Irequencies ..................................... 4.10
5. REFERENCES
Table oI contents
MANUAL FOR RISK ANALYSIS OF TRANSPORTATION ACTIVITIES, PREFACE
This manual Ior risk analysis oI transportation activities gives the state-oI-the-art, and is based
on analyses oI accident reports and agreements on parameter values between all parties
involved, both governmental authorities and transport operators. Agreement has been reached
through a process Ior which the Dutch Ministry oI Transport has Iinal responsibility.
The application oI this part oI the manual and its rules Ior the execution oI risk studies is
restricted to the transport oI dangerous goods on public routes Ior vehicles, trains and water
transport, and to pipelines outside the boundary limits oI plants and terminals, and shunting
yards owned and operated by railway operators such as Dutch Railways.
The models and data within this part oI the manual are speciIic Ior transportation activities
outside plants and terminals. They are thereIore not to be applied in risk analyses Ior
transportation activities and/or stationary equipment within plants and terminals.
PreIace
1. SELECTION OF RELEVANT ROUTE SECTIONS
1.1 Introduction
In order to evaluate whether the risk oI transporting hazardous substances along a speciIic route
complies with external saIety criteria, the individual risk and the societal risk should be
calculated. Eortunately it is not always necessary to perIorm a detailed, time-consuming and
costly QRA. Insight into the level oI risk Ior a speciIic route section may be gained by applying
the Iollowing three steps, which are increasingly detailed:
1. Comparing the Irequencies oI annual transport movements with threshold values gives a Iirst
quick scan oI the risk levels. When the annual Irequency oI transport movements along the
route is less than the threshold value a quantiIication oI the external saIety risks is not
needed. In such cases no external saIety problem Iormally exists although, oI course,
incidents with release oI hazardous materials still may occur. Whenever the threshold values
are exceeded or are not applicable Ior the speciIic situation, a quantiIication oI the risk
should be made.
2. At Iirst the risk may be relatively simply estimated by using IPORBM (chapter 2). IPORBM
quantiIies the risk in terms oI individual risk and societal risk and comparison to the Dutch
external saIety criteria is straightIorward. Although IPORBM usually produces results with
adequate precision, speciIic situations may require an even more accurate and detailed risk
calculation as described in chapter 3. Currently no general guidelines can be given as to
whether an IPORBM calculation is suIIicient, or whether a more detailed analysis is
required. Complex situations diIIering Irom the simple standard situation assumed in
IPORBM should be treated with care. In particular, a more detailed analysis will be
necessary in cases where the calculated results approximate to the speciIied risk criteria, or
where the outcome oI a calculation is oI crucial importance to the saIety issue involved and
will thereIore be heavily debated. In these cases a speciIic risk calculation should be made.
3. PerIorming a detailed QRA, as outlined in chapter 3.
In the Iollowing paragraphs the threshold values Ior the Irequency oI transport movements on
roads (paragraph 1.2), railways (paragraph 1.3) and inland waterways (paragraph 1.4) are given.
These thresholds provide a quick selection oI those situations where Iurther consideration and
quantiIication oI the external saIety risks are appropriate. The thresholds given are conservative
in that they ensure that every situation that might be relevant with respect to non-compliance to
the Dutch external saIety criteria will be considered. Thresholds Ior the transport oI hazardous
materials by pipeline are described in |VNG98|.
1.2 Threshold values for roads
It appears that the external risk oI road transport in the Netherlands is strongly dominated by the
transport oI LPG |AVIV97|. The threshold values are thereIore given in terms oI the annual
Irequency oI transport movements oI the substance category GE3 which are Ilammable gases
like LPG. The Iollowing road typology is used:
Motorways
Non-urban roads
Urban roads
Selection oI relevant route sections 1.1
1.2.1 Individual risk
Table 1.1 gives the threshold values Ior annual Irequencies Ior LPG transport movements and
Ior all hazardous substances below which no 10
-6
individual risk contour exists.
1able 1.1 1hreshold values below which no 10
-6
individual risk contour exists
Road type Threshold for LPG Threshold for all hazardous substances
(movements/year) (movements/year)
Motorway 6500 27000
Non-urban 2300 7500
Urban 8000 22000
Notes:
1. The thresholds should be checked Iirst Ior LPG and then Ior all hazardous substances.
2. The thresholds apply to an open road situation (no obstacles) and average road saIety. When
experience indicates that the local accident rate is increased, such as at a level crossing, the
thresholds should be applied with caution and a more detailed quantiIication oI the risks
should be made.
3. 'All hazardous substances means substances categorised in one oI the substance categories:
Ilammable liquid (LE), Ilammable gas (GE), toxic liquid (LT) or toxic gas (GT) according to
|AVIV95|.
4. The threshold values given Ior all hazardous substances apply to a broad range oI
combinations oI transported substances, except when there are signiIicant numbers oI toxic
substances being transported. Then the thresholds are not applicable and a more detailed
quantiIication oI the risks should be made. Eor guidance, the annual Irequencies required tot
give an individual risk exceeding 10
-6
per year Ior toxic substances is given in Table 1.2.
1able l.2 1hreshold values required to give an individual risk exceeding 10
-6
per year
Movements/year
Substance category Motorway Non-urban Urban
GT2 or GT3 8000 3000 10000
GT4 or GT5 4000 2000 8000
LT2 10000 3000 8000
LT3 2000 700 2000
LT4 700 300 800
1.2.2 Societal risk
The societal risk is also dominated by the transport oI liqueIied, pressurised, Ilammable gases,
mainly LPG. The societal risk depends on the annual Irequency oI transport movements and on
the distance and density oI the population along the road. Table 1.3 gives the threshold below
which the annual Irequency oI LPG transport movements Ior a particular population density
along a speciIied route does not lead to a societal risk that exceeds the societal risk criterion.
Thresholds Ior all hazardous substances are given in Table 1.4.
1able 1.3 1hreshold values below which the societal risk criterion is not exceeded
Population density Threshold for LPG (movements/year)
(pers/ha) (one-sided development)
Motorway Non-urban Urban
100 500 200 500
90 600 200 700
80 700 200 800
70 900 300 1100
60 1300 400 1500
50 1800 600 2000
40 2800 1000 3500
30 5100 1800 6000
20 11000 4000 13500
10 45500 16000 53000
Population density Threshold for all hazardous substances (movements/year)
(pers/ha) (one-sided development)
Motorway Non-urban Urban
100 2500 900 3500
90 3500 1200 4000
80 4000 1500 5000
70 5500 2000 6500
60 7500 2500 9000
50 10500 4000 13000
40 16500 6000 20500
30 29500 10500 36500
20 66500 23500 82000
10 266000 94000 326000
Notes:
1. The thresholds should be checked Iirst Ior LPG and then Ior all hazardous substances.
2. The thresholds are conservatively Iormulated i.e. the development (population) is assumed to
be directly adjacent to the road.
3. The population density is determined as the average density within 200 meters Irom the road
edge. When the maximum density is more than three times the average, the maximum value
should be used instead.
4. When the development (population) is on both sides oI the road, the Irequencies should be
divided by a Iactor oI 4.
5. The thresholds apply to an open road situation oI average road saIety. When local experience
indicates that the accident rate may be higher, such as at a level crossing, the thresholds
should be applied with caution. A more detailed quantiIication oI the risks using IPORBM
should be made.
6. The threshold values given apply to a broad range oI combinations oI transported substances.
Only when very toxic substances are being transported, such as the categories LT3, LT4, or
GT5, the thresholds are not applicable and a detailed quantiIication oI the risks using
IPORBM should be made.
1.3 Threshold values for railways
The external risk oI the transport oI hazardous substances by rail is dependent on the substances
transported and the track characteristics, notably train speed. The Iollowing typology oI railway
tracks is used:
High speed tracks (>40 km/hr)
Low speed tracks (40 km/hr).

The substances are categorised in a scheme that is (Ior historical reasons) speciIic Ior railway
transport. This categorisation is solely based on the Kemler code oI the substance.
The categories Ior railway transport are given in Table 1.5.
1able 1.5 Substance categories for railway transport
Substance category Kemler code Note
A 23, 263, 239 Elammable gas, liqueIied under
pressure, such as LPG
B2 26, 265, 268 (chlorine excluded) Toxic gas, liqueIied under
pressure, such as ammonia
B3 chlorine itselI Extremely toxic gas, liqueIied
under pressure, like chlorine
Chlorine chlorine itselI In dedicated chlorine trains
D3 Acrylonitrile itselI Toxic liquid, such as acrylonitrile
D4 66, 663, 668, 886, X88, X886 Extremely toxic liquid, such as
hydrogen Iluoride
C3 33, 336 (Acrylonitrile excluded), Extremely Ilammable liquid, such
338, 339, X323, X333, X338 as motor spirit
Table 1.6 gives the threshold annual values Ior C3 transport movements and Ior all hazardous
substances below which no 10
-6
individual risk contour exists.
1able 1.6 1hreshold annual values below which no 10
-6
TRACK TYPE
High speed Low speed
Threshold Ior C3 3000 10
-6
not exceeded
(rail cars/year)
Threshold Ior all hazardous 7000 10
-6
not exceeded
substances (rail cars/year)
Notes:
1. The thresholds should be checked Iirst Ior C3 and then Ior all hazardous substances.
2. The thresholds apply to moving rail cars and open track situations (no tunnels, no obstacles).
Eor shunting yards and private sidings a speciIic calculation methodology is available (see
chapter 3.3)
3. Eor low speed tracks more than 55,000 transport movements annually are required Ior an
individual risk exceeding 10
-6
per year. Eor Dutch standards this is unrealistically high.
ThereIore thresholds Ior low speed tracks are not presented.
4. 'All hazardous substances means substances categorised in one oI the substance categories
speciIied in Table 1.5.
5. The threshold values given apply to a broad range oI combinations oI transported substances.
However, when signiIicant numbers oI toxic liquids in the category D3 or D4 are being
transported the thresholds are not applicable and a more detailed quantiIication oI the risks
should be made. Eor guidance, the annual Irequencies needed Ior high speed tracks to
generate an individual risk higher than 10
-6
per year Ior category D3 and D4 are 13,000 and
9,000 respectively.
1.3.2 Societal risk
The societal risk depends upon the annual Irequency oI transport movements and on the distance
and density oI the population along the track. The level oI the societal risk is strongly dependent
on the presence oI extremely toxic pressurised gases.
The societal risk criterion may be exceeded when the annual Irequency oI rail cars exceeds the
values mentioned in Table 1.7.
1able 1.7 1hreshold values for transport of toxic gases below which the societal risk criterion
is not exceeded
Track type
Substance category High speed Low speed
B3 (rail cars/year) 60 2000
Chlorine (rail 300 8000
cars/year)
Table 1.8 gives the thresholds below which the annual Irequency oI LPG transport movements
Ior the population density along the speciIied route does not lead to a societal risk that exceeds
the societal risk criterion. Thresholds Ior all hazardous substances are also given.
Population density Threshold for LPG Threshold for all hazardous
(pers/ha) (rail cars/year) substances (rail cars/year)
High speed Low speed High speed Low speed
track track track track
100 1600 8000 7500 37500
90 2000 10000 9000 46000
80 2500 12500 12000 58500
70 3000 16000 15000 76500
60 4500 22000 21000 104000
50 6500 32000 30000 150000
40 10000 50000 47000 234000
30 20000 88000 83000 416000
20 40000 200000 187000
Notes:
1. The thresholds should be checked Iirst Ior the extremely toxic gases, then Ior LPG, then Ior
all hazardous substances.
2. The thresholds are conservatively Iormulated i.e. the development (population) is assumed to
be directly adjacent to the track.
3. The population density is determined as the average density within 200 meters Irom the track.
When the maximum density is more than three times the average, the maximum value should
be used instead.
4. Table 1.8 is Ior one-sided developments oI the area along the track. When the development
(population) is on both sides oI the track, the Irequencies should be divided by a Iactor oI 4.
5. The thresholds apply to moving rail cars and open track situations (no tunnels, no obstacles).
Eor shunting yards and private sidings a speciIic calculation methodology is available (see
chapter 3.3).
6. 'All hazardous substances means substances categorised in one oI the substance categories
speciIied in Table 1.5.
1.4 Threshold values for inland waterways
Inland waterways are characterised by navigability class according to the CEMT convention.
The navigability class deIines a maximum size oI vessel in relation to the dimensions oI the
waterway. The main waterways Iall into the CEMT classes 4, 5 and 6. Navigability classes Ior
speciIic waterways are given in Table 1.9 and may also be Iound in |AVV97a|.
1able l.9 Navigability classes for specific waterways
Waterway Class Waterway Class
Eemskanaal 5 Waal 6
V Starckenborghkanaal 5 Beneden Merwede 6
Prinses Margrietkanaal 5 Noord 6
IJssel 5 Hollandsch Diep 6
Nederrijn 5 Schelde-Rijnkanaal 6
Lek 5 Maas 5
Amsterdam-Rijnkanaal 6 Julianakanaal 5
Nieuwe Maas 6 Kanaal Gent-Terneuzen 6
Oude Maas 6 Hollandsche Ijssel 5
Oude Rijn/Gouwe 4
The accident rate depends on the navigability class. However, the accident rate may vary by
several orders oI magnitude between sections oI the same waterway due to the presence oI local
Iactors. The use oI thresholds Ior the annual transport movements is thereIore more problematic
than Ior the other modes oI transport. The thresholds given should thereIore be used with
extreme caution, and one should be attentive to the presence oI Iactors indicative oI an increased
local accident rate.
Transport oI hazardous substances on the class 4 and 5 waterways consists mainly oI Ilammable
liquids. On class 5 and 6 waterways toxic substances are also encountered.
The individual risk near the waterway is dominated by the transportation oI Ilammable liquids.
Threshold values Ior the Irequency oI annual transport movements Ior substance category LE2,
below which no 10
-6
individual risk contour exists, are given in Table 1.10.
1able 1.10 1hreshold values below which no 10
-6
Navigability class Threshold for substance category LF2
(movements/year)
4 7000
5 6500
6 3000
Notes:
1. The thresholds apply to a situation oI average traIIic saIety. When local nautical experience
indicates that the accident rate may increase due to the presence oI a bend with a limited
view, or a dock or harbour entrance, or other such Iactors, the threshold values given should
be applied with extreme caution. A more detailed quantiIication oI the risks using IPORBM
should be made. The same applies to situations in which the distribution oI traIIic over the
width oI the waterway deviates signiIicantly Irom a uniIorm one.
1.4.2 Societal risk
Societal risk depends upon the annual Irequency oI transport movements and on the distance
and density oI the population along the waterway. lt appears that the societal risk in the
Netherlands is dominated by the transport oI toxic substances. The thresholds are thereIore
Iormulated in two instances:
1. When extremely toxic liquids LT3 or LT4 are transported, the societal risk should always be
quantiIied.
2. Table 1.11 gives the threshold values Ior the annual transport movements oI ammonia
(liqueIied under pressure) which, in combination with the given population density along the
route, does not lead to a societal risk that exceeds the societal risk criterion.
Population density Threshold for liquefied ammonia under pressure
(pers/ha) (movements/year)
(one-sided development)
Navigability-class Navigability-class Navigability-class
4 5 6
100 6000 4500 2000
90 7500 5500 2500
80 9500 7000 3000
70 12000 9000 4000
60 5500
50 8000
40 12000
30 22000
Notes:
1. The threshold values apply to a situation oI average traIIic saIety. When local nautical
experience indicates that the accident rate may increase due to the presence oI a bend with a
limited view, or a dock or harbour entrance, or other such Iactors, the threshold values given
should be applied with extreme caution. A more detailed quantiIication oI the risks should
be made. This also applies to situations in which the distribution oI traIIic over the width oI
the waterway deviates signiIicantly Irom a uniIorm one.
2. When the development (population) is on both sides oI the waterway, the Irequencies should
be divided by a Iactor oI 2.
3. II ammonia is transported in semi-pressurised tankers (temperature below 278 K) no
threshold values apply. The societal risk does not exceed the societal risk criterion.
2. IPORBM
2.1 Introduction
IPORBM is the Dutch acronym Ior Inter Province Committee Ior Risk Calculation
Methodology. AVIV BV consultants have developed the IPORBM soItware program. It is a
standardised calculation methodology Ior determining the external risks involved when
transporting hazardous substances (Ilammable and toxic gases and liquids in bulk) by road,
railway, waterway or pipeline |IPORBM|. This risk calculation methodology oIIers a means Ior
quickly calculating the risk level along a deIined transport route, based on a limited amount oI
input data. A method to identiIy the transport routes and speciIic locations where a risk
calculation should be made is described in chapter 3. In this chapter a description oI the
program IPORBM and the required input data Ior each transport modality will be given.
Sources where the required input data, regarding the transport systems, transport streams,
accident Irequencies and population densities may be obtained are described in chapter 4.
Although IPORBM usually produces results with adequate precision, speciIic situations may
require an even more accurate and detailed risk calculation (QRA) described in chapter 3. As it
is, no general guidelines can be given as to whether an IPORBM calculation is suIIicient, or a
more detailed analysis is required. SpeciIic situations, which diIIer Irom the standard situation
assumed in IPORBM and described below, should be treated with care. In particular, a more
detailed analysis will be necessary in cases where the calculation results approximate to a
certain predetermined criterion, or where the outcome oI a calculation is oI crucial importance
to the saIety issue involved and will thereIore be heavily debated. Some oI the deIault values oI
the parameters in the program may be changed to meet the requirements oI a speciIic situation.
2.2 IPORBM: a general description
The program is set up to determine the societal risk and the individual risk resulting Irom the
transport oI hazardous substance categories along a certain route, called a 'trajectory (route
section). In IPORBM all 'trajectories (route sections) are deIined as straight lines and it is
assumed that no obstacles such as tunnels, or sound barriers, are present (open route). Eor
certain speciIic situations, Ior example iI obstacles are present, the route is elevated, or there are
sharp bends in the route, IPORBM will not give adequate results and a more detailed calculation
may be appropriate.
In IPORBM the diversity oI substances transported is reduced to a smaller range oI substance
categories by using standard categories containing substances with similar risk Iactors. Each
substance category is characterised by a speciIic example substance. This categorisation oI
substances is outlined in chapter 4 and, amongst other things, is based on aggregation, volatility,
Ilammability and toxicity. Eor pipelines, risks can only be calculated in IPORBM Ior a limited
number oI substances, and combinations oI pipeline diameter and pipeline operating pressures.
In IPORBM the transport risks oI every transport mode are characterised by a limited number oI
representative loss oI containment events (LOCs) and a limited number oI typical transport units
and transport route characteristics. These LOCs have been derived Irom a number oI reports and
IPORBM 2.1
are described in detail in chapter 3 and in |IPORBM|. In IPORBM a minor and a major LOC is
usually deIined. The physical eIIects oI these LOCs and resulting events (scenarios) have been
calculated beIorehand in IPORBM Ior the six Iixed weather types given in Table 2.1.
1able 2.1 Weather types used in IPORBM
Weather class Daytime Night-time
(Pasquill class, wind speed) probability probability
B3.0 0.220 0.000
D1.5 0.122 0.149
D5.0 0.299 0.262
D9.0 0.359 0.261
E5.0 0.000 0.112
E1.5 0.000 0.216
Eixed parameters were used characterising the surroundings oI the transport route, the transport
unit, and the release and dispersion oI the hazardous substance. The results oI these calculations
have been entered into a result matrix. In this way, the program only has to perIorm a limited
number oI calculations Ior a complete risk evaluation. The user has no Iurther control over
accident scenarios or eIIect calculations. The user oI the program merely enters a description oI
the route section, the type oI route section (route characteristics), the population data, the annual
Irequency oI transport movements per substance and transport category, ignition probabilities
and the probabilities oI the six weather types, Irom which calculations are made. Wind direction
cannot be taken into account because a uniIorm probability is assumed in IPORBM. The use oI
speciIic accident Irequencies is recommended. II deIault values Ior the accident Irequencies are
used, it should be ascertained whether this leads to an over- or under-estimation oI the risks.
Population density is indicated in IPORBM by rectangles along the route, with a uniIorm
population density per rectangle. Input parameters include the size oI these rectangles, the
distance Irom either side oI the route (measured Irom the axis oI the transport route), and the
average population density per rectangle. Situations where large groups oI people are present
during short periods oI time, as in stadiums, should be addressed in a speciIic calculation.
The population density and the transport movements oIten depend on the time oI the day. The
population density can be divided into day and night Iractions. Also, the Iractions oI the
transport movements that take place during the day and night can be speciIied. Chapter 3
outlines how, and to what level oI detail, population densities must be speciIied.
In the case oI Ilammable gases and liquids, an immediate and a delayed ignition probability are
deIined Ior each minor and major outIlow (LOC). DeIault values Ior each transport mode are
given in the next sections. In individual risk calculations, the gas cloud is assumed to ignite at
its maximum size. In societal risk calculations, the time oI ignition depends on the presence oI
ignition sources. The maximum probability oI a delayed ignition oI the gas cloud is equal to the
value entered. Eor delayed ignition calculations two options are available. The user can deIine
whether the ignition probability is dependent on the population density under the developing gas
cloud or is also a Iunction oI the travel time oI the gas cloud. The Iirst option is recommended.
IPORBM 2.2
Situations may occur where other types oI ignition sources, such as traIIic and industry related
sources, are present and the approximation used in IPORBM Ialls short. A more appropriate
detailed calculation should be considered.
AIter entering all the required input data and the program has perIormed the calculation, the
results are displayed as individual risk contours along the total transport route, and as societal
risk graphs per kilometre section. Input data and calculation results can be saved to disk.
Calculation results can be displayed on the screen as text, as societal risk graphs (also known as
EN curves), and as individual risk contours. Graphical representations oI the results can be
printed using a printer or a plotter. The contribution oI the diIIerent substance categories to the
overall risks can be shown.
2.3 Road
The required input data Ior road transport are visualised in Eigure 2.1 and Eigure 2.2. In
IPORBM, the road transport system is described using a predeIined road type, the length oI the
road section, and the annual Irequency oI movements oI Iully loaded transport Ior categories oI
substances (combination oI transported substances: LE-GT) and the population data along the
transport route. The daytime Iraction oI the total transport movements can also be deIined. This
ratio indicates the transport Iraction that takes place during the meteorological day. Eor road
transport, this Iraction has a deIault value oI 0.8 (80). The probability oI an outIlow oI a
hazardous substance oI a certain magnitude is dependent on the type oI road and the type oI
transport unit. IPORBM contains deIault outIlow Irequency values Ior pressurised tankers and
atmospheric tankers, Ior each oI Iour road types. These types are motorways, outside built-up
area, inside built-up area, and generic. The user can, however, supply alternative outIlow
Irequencies Ior atmospheric tankers Ior speciIic route sections (Ior pressurised tankers the
outIlow Irequencies are determined by IPORBM, based on a Iixed ratio between atmospheric
and pressurised tankers). The procedure Ior determining location speciIic outIlow Irequencies
using location speciIic injury accident Irequencies is outlined in chapter 4. DeIault ignition
probabilities are shown in Eigure 2.2. In the case oI a pool Iire the delayed ignition eIIects are
assumed to be equal, and the immediate and delayed ignition probabilities are added. The
deIault values are recommended, although the deIault values Ior the ignition probabilities may
be changed.
IPORBM 2.3
Figure 2.1 1he road modality main entry screen
Figure 2.2 1he Road Parameters dialog box
2.4 Rail
The Dutch railway system may be seen as an assembly oI shunting yards, open tracks and
private sidings (chapter 3.3). IPORBM addresses the risks associated with the transport by rail
oI hazardous substances, Ilammable and toxic gases and liquids in bulk, in moving cars Ior open
track situations. However, the user should bear in mind that accidents related to the arrival and
departure oI trains (to be) shunted at yards, to and Irom open tracks, are not considered in
IPORBM 2.4
IPORBM. ThereIore, Ior locations in the proximity oI shunting yards, a detailed QRA has to be
made.
In IPORBM, the rail transport system is described by the railway route section type, the length
oI this section, the type oI train in which the hazardous substances are transported, the annual
Irequency oI transport movements Ior each substance category and type oI train, and the
population data along the route. The required input data Ior rail transport are visualised in
Eigure 2.3 and Eigure 2.4.
Six substance categories (A, B2, B3, C3, D3 and D4) are deIined, based on the Kemler code
(Danger Index code or GEVI code) oI the substances carried (see Table 1.5). Data on the
transport oI hazardous substances may be gathered Irom inIormation supplied about the Dutch
railways (NS), the main operator in the Netherlands. By deIault, 33 oI rail transport takes
place during the day, and 67 oI rail transport occurs at night.
Two railway types, high speed and low speed, are distinguished in respect oI the diIIerence in
initial accident probability. The average initial accident probability Ior a rail car on an open track
is 3.6 * 10
-8
per rail car kilometre. The average number oI rail cars involved in an accident
depends on the speed oI the train. Taking this aspect into account results in average accident
probabilities oI 2.2 * 10
-8
Ior speeds below 40 km/h (low speed) and 4.5 * 10
-8
Ior speeds
exceeding 40 km/h (high speed) per rail car kilometre.
Three types oI trains are distinguished in IPORBM; mixed trains, block trains and dedicated
chlorine trains. Eor dedicated chlorine trains, which have only rail cars with chlorine, a IiveIold
lower initial accident probability is assumed due to the Iact that these trains are subjected to
extra stringent saIety precautions in the Netherlands. Mixed trains are trains with both
Ilammable gas (substance category A) rail cars and Ilammable liquids (substance category C3)
rail cars. Eor mixed trains, the assumed BLEVE event probability Ior Ilammable gases is twice
that Ior block trains, thus compensating Ior the increased Iire hazard resulting Irom cars
containing Ilammable liquids. Block trains (Ior Ilammable gas) are trains without any rail cars
containing Ilammable liquids.
In IPORBM, deIault standard average accident Irequencies are used, which depend on railway
type and train type as described above. The average situation is deIined as an open track with an
average number oI crossings (0.66 per kilometre) and points (0.27 per kilometre) and
saIeguarded with an ATB saIety system. However, speciIic situations may be identiIied where
the use oI a speciIic Irequency Ior a part oI a route section is appropriate. The average accident
Irequency oI 3.6 * 10
-8
per rail car kilometre may be multiplied by a Iactor, which is given in
Table 2.2 Ior some speciIic situations. No correction Iactors have yet been derived Ior the
speciIic situation oI trains passing through stations and shunting yards.
IPORBM 2.5
IPORBM 2.6
1able 2.2 Factors for the multiplication of the average accident-frequency 3.6 ' 10
-8
per rail
car kilometre for specific situations
Specific situation Factor
High speed (>40 km/h) 1.26
Low speed (40 km/h) 0.62
ATB-new generation 0.9
No crossings and no points (/kilometre) 0.6
One -crossing, no point (/kilometre) 0.8
One point, no crossing (/kilometre) 1.5
A crossing and a point (/kilometre) 1.8
Hot box detection 0.8
DeIault ignition probabilities Ior Ilammable gases (A) and liquids (C3) are shown in Eigure 2.4.
In the case oI a pool Iire, the immediate and delayed ignition eIIects are assumed to be equal
and the probabilities are added. The deIault values are recommended, although the deIault
values Ior the ignition probabilities may be changed.
Figure 2.3 1he railway modality main entry screen.
Figure 2.4 1he Railway Parameters dialog box
IPORBM 2.7
2.5 Inland waterway
The inland waterway system is described in IPORBM by the type and width oI the waterway,
the length oI the waterway section, the annual Irequency oI transport movements per substance
category and vessel type, and data on population along the waterway. The required input data
are shown in Eigure 2.5 and Eigure 2.6.
In IPORBM, deIault values are used Ior the initial accident probability Ior each type oI
waterway. Waterways are characterised by navigability class. Navigability class 6, Ior example,
describes a major waterway. InIormation on all Dutch waterways is given in |AVV97a|. The
navigability class oI a number oI Dutch waterways is given in chapter 1.4. Initial accident
probabilities are deIined per part oI a waterway route section, and reIer to accidents resulting in
major damage to one or more vessels. The program then uses this inIormation, together with the
type oI waterway and vessel speciIied, to calculate the probability oI LOC. Accidents can be
assumed to be distributed across the width oI the waterway. In IPORBM the waterway is
thereIore divided into nine lanes oI equal width, and Ior each lane a Iraction oI the accident
Irequency must be deIined. The deIault initial accident Irequencies used in IPORBM are
indicative; they can only be used Ior a Iirst rough approximation oI risk levels. There is a wide
spread in initial accident probabilities. Near manmade objects in the waterway (bridges, locks),
or near crossings and narrow bends, the accident Irequencies can increase by ten or even a
hundred times. Eor these speciIic situations the use oI a speciIic Irequency, based on nautical
judgement and historical data, is appropriate.
Data on the nature and volume oI transport are oIten obtained Irom registered lock passages. Eor
waterway transport, 8 substance categories and Iour vessel types may be deIined as shown in
Table 2.3. The vessel types are: pressure tankers (PRESSURE), semi-pressurised tankers
(SEMI), single walled and double walled tankers. IPORBM does not addresses the risks
associated with seagoing vessels. By deIault, 50 oI waterway transport takes place during the
daytime, and 50 oI waterway transport occurs at night. DeIault ignition probabilities Ior
Ilammable gases (GE3) and liquids (LE 1, LE2) are shown in Eigure 2.6. In the case oI a pool
Iire, the immediate and delayed ignition eIIects are assumed to be equal and the probabilities are
added. The deIault values are recommended, although the deIault values Ior the ignition
probabilities may be changed.
1able 2.3 Substance categories and vessel types
Main substance category Subcategory Vessel type
Elammable gas GE3 Pressure
Toxic gas GT3 Pressure
Toxic gas GT3 Semi-pressurised
Elammable liquid LE1 Single/double walled
Elammable liquid LE2 Single/double walled
Toxic liquid LT1 Single/double walled
IPORBM 2.8
Figure 2.5 1he waterway modality main entry screen
Figure 2.6 1he Waterway Parameters dialog box
IPORBM 2.9
2.6 Pipeline
The pipeline system is described in IPORBM by substance, type oI pipeline, length, diameter,
operating pressure and data on population along the pipeline route. IPORBM is applicable Ior
the substances, diameters and operating pressures shown in Table 2.4. II other substances are
being transported, then IPORBM cannot be used and the risks will have to be calculated in
another way. In the Netherlands, Ior several substances such as natural gas and K1 and K2
liquids, directives in the Iorm oI (zoning) regulations apply. Eor these substances no individual
risk calculation can be made; instead, the required zoning distances shown in chapter 3.5 are
given by IPORBM. The societal risk can be calculated Ior all substances mentioned in Table 2.4.
Unlike the other modalities, risk calculations Ior pipelines are restricted to the risks resulting
Irom the operation oI a single pipeline. This approach is in accordance with current policy in the
Netherlands. The required input data Ior IPORBM are visualised in Eigure 2.7 and Eigure 2.8.
DeIault ignition probabilities Ior Ilammable gases and liquids are shown in Eigure 2.8. The user
may change the deIault values. In the individual risk calculation the assumption is made that any
Ilammable gas released will always be ignited.
InIormation on the nature and mode oI hazardous substance transport by pipeline can be
obtained Irom the pipeline operator. It is assumed that pipeline transport takes place around the
clock, that is 50 during daytime and 50 during night-time. Accident Irequencies have to be
entered Ior each part oI a route section. DeIault values are given Ior each type oI pipeline. The
Iollowing types are distinguished:
Pipeline on dedicated route Ior pipelines
HTL: pipeline belonging to the natural gas high pressure pipeline network
NEN 3650/RTL new: pipeline built in compliance with NEN 3650
NEN 3650/RTL old: pipeline not built in compliance with NEN 3650
DeIault values may be changed whenever a speciIic saIety management system or extra, above
standard, saIety measures are taken to minimise outIlow (automatic block-in valves) or to
protect the pipe Irom external interIerence, such as extra ground cover or using double walled
pipes. However no speciIic recommendations Ior the correction Iactors to be applied can be
given.
1able 2.4 IPORBM pipeline types
Substance category Diameter inches Operating pressure bar
Natural gas 2, 4, 6, 8, 10, 12, 14, 16, 18, 24, 40, 60, 90
30, 36, 42, 48
Ammonia 4, 6, 8 10, 14, 20
Chlorine 2, 3, 4 12 through 20
Ethylene (Ethene) 6, 8, 10 50, 75, 100
Ethylene oxide 6 through 10 5 through 10
K1 4, 6, 8, 10, 12 all pressures
K2 4, 6, 8, 10, 12, 14, 16, 18, 24, all pressures
30, 36
Carbon monoxide 6, 8, 24 20, 30
Propane 6, 8, 10 50, 75, 100
Vinyl chloride 8, 10 50, 75, 100
IPORBM 2.10
Figure 2.7 1he pipeline modality main entry screen
Figure 2.8 1he Pipeline Parameters dialog box
IPORBM 2.11
3. DETAILED QRA
3.1 Introduction
This chapter gives an outline oI a detailed QRA Ior the transport oI dangerous substances in
bulk based on current practice in the Netherlands. The loss oI containment events (LOCs) that
need to be included in the QRA are identiIied. The corresponding Iailure Irequencies and the
various events Iollowing a release, and their probabilities, are given. The modelling oI the
source term, dispersion, exposure and damage, and the calculation and presentation oI the
results, however, are not described in detail. Only the diIIerences in modelling to that described
Ior stationary installations at an establishment are highlighted.
The QRA rules given are general and to be applied to open situations, i.e. transport routes where
no obstacles, such as tunnels and noise barriers, are present. Situations in which obstacles are
present diIIer in several aspects Irom an open situation and thereIore need to be considered on a
case by case basis. In paragraph 3.2.5 these aspects are qualitatively described.
The basic data needed in conducting a QRA Ior a speciIic transportation route include the:
Description oI the transport streams (number oI yearly loaded transport units per substance or
category, during daytime and night-time)
Description oI the transport units (characteristic inventory)
Description oI the transport route (road type, obstacles present)
Description oI the number oI accidents and degree oI traIIic in order to determine accident
Irequencies
Description oI the ignition sources
Properties oI transported (representative) substances
Terrain classiIication oI the surroundings oI the transport route
Meteorological data
Population present in the surroundings oI the transportation route

Guidelines Ior obtaining these data, recommended values and reIerences to inIormation sources
are given in the next paragraphs and in chapter 4.
The level oI detail oI the QRA may be varied depending on whether or not location speciIic
accident Irequencies are used or speciIic hazardous substances or substance categories are
included in the QRA. PreIerably, location speciIic accident Irequencies should be used in the
QRA. Eurthermore, on a case by case basis, the deIault release scenarios and events may be
altered to meet the requirements oI the speciIic situation. No general rules can be given. A
method to identiIy transportation routes and speciIic locations where a QRA should be made is
described in chapter 3. In chapter 2 IPORBM, a standardised but simpliIied calculation
methodology Ior determining the risks involved in transporting hazardous substances, is
described.
Detailed QRA 3.1
3.2 Road transport
3.2.1 Loss of containment events
The risks oI the transport oI dangerous substances by road are mainly determined by transport
in bulk. The transport oI small packages (drums, cylinders) and the transport oI explosives and
radioactive materials are hitherto not considered in a QRA. In the QRA a distinction is made
between Ilammable and toxic substances (liquids and gases) transported in atmospheric and
pressurised transport units (tankers and tank containers).
The typical inventory oI an atmospheric tanker in the Netherlands is 23 metric tonnes. Eor a
pressurised tanker with Ilammable gas a typical inventory is 20-25 metric tonnes and Ior a
pressurised tanker with toxic gas a typical inventory oI 16 metric tonnes may be assumed.
The LOCs Ior road transport are Irom |AVIV94, VeVoWeg96| and are given below.
Atmospheric tankers and tank containers:
release oI the complete inventory.
release oI 5 m
3
oI the inventory.
release oI 0.5 m
3
oI the inventory.
Pressurised tankers and tank containers:
instantaneous release oI the complete inventory oI the transport unit.
continuous release Irom a hole with an eIIective diameter oI 50 mm (2 inches).

Notes:
1. No distinction has to be made in the QRA between tank containers and tankers.
2. Some substances such as hydrogen are transported in bulk as compressed gases or as
reIrigerated liquids. As the Irequency oI yearly transport movements is small compared to the
Irequency oI transport movements oI substances that determine the risks, these substances
may in most cases oI practical interest be omitted in the calculation.
3. LOCs Ior transport oI explosives and radioactive materials diIIer Irom those given Ior
Ilammable and toxic substances (liquids and gases). As explosives and radioactive materials
are rarely transported, they need not be included in a QRA.
4. A release oI 0.5 m
3
Irom an atmospheric transport unit will result in a small pool. In most
cases oI practical interest, such as open road situations, this LOC may be omitted in the
calculation.
3.2.2 Accident and outflow frequency, events and event probabilities
The Irequency oI a hazardous event at a speciIic location on the transport route per transport unit
per year, may be seen as the product oI the initial accident Irequency (per unit and km), the
Detailed QRA 3.2
probability oI a signiIicant release (> 100 kg) given an accident Ior the speciIic transport unit
and the probability oI the hazardous event given a signiIicant release.
The product oI the initial accident Irequency and the release probability is called the outIlow
Irequency. Generic values (per transport unit km) Ior diIIerent road types are derived in
|AVIV94| and given in Table 3.1 Ior pressurised and atmospheric transport units. The generic
values are based on the number oI known hazardous substance accidents with a known release in
a certain period, and data on the transport oI hazardous substances in the same period, estimated
Irom transport statistics. The probability oI signiIicant release (> 100 kg) given an initial
accident cannot be deduced directly, because the initial accident Irequency oI transport units
with dangerous substances (including accidents with no release) as yet cannot be derived Irom
general accident statistics.
1able 3.1 Outflow frequencies for different road types
Road Type Outflow frequency /veh.km
Pressurised Atmospheric
Motorway 4.32 * 10
-9
8.38 *10
-9
Outside built-up area 1.22 * 10
-8
2.77 *10
-8
Inside built-up area 3.54 * 10
-9
1.24 * 10
-8
In a detailed, location speciIic QRA it is current practice in the Netherlands to assume that the
outIlow Irequency is a linear Iunction oI the injury accident Irequency. The outIlow Irequency
Ior a particular road section may then be determined by multiplying with the ratio oI the
location speciIic accident Irequency to the average injury accident Irequency Ior that road type.
This procedure is outlined in section 4.3.2.
The hazardous events Iollowing a release to be considered in the QRA include:
Toxic exposure
BLEVE
Jet Iire
Pool Iire
Elash Iire
Explosion
Eor the QRA deIault, immediate ignition probabilities are used as given in Table 3.2. Delayed
ignition in Societal Risk calculations should be modelled as a Iunction oI the distribution oI
ignition sources (detailed QRA Ior existing/planned situation). When the speciIic locations oI
ignition sources are unknown the values Irom Table 3.2 may be used instead (QRA Ior a generic
situation). Delayed ignition in an Individual Risk calculation should be modelled to give the
maximum eIIects. Ignition at maximum cloud area should be assumed.
Detailed QRA 3.3
1able 3.2 Ignition probability
Substance category Ignition probability
Immediate Delayed
Elammable liquid, category LE2 0.065 0.065
Elammable liquid, category LE1 0.0043 -
Elammable Gases 0.8 0.2
Events and deIault event probabilities Ior Ilammable substances are given in Eigure 3.1 and
Eigure 3.2. The events and eIIects Iollowing immediate and delayed ignition oI a release oI
Ilammable liquids category LE2 are almost the same, so the probabilities oI delayed ignition is
added to the probability oI immediate ignition as shown in Eigure 3.2. Delayed ignition oI a
release oI Ilammable liquids category LE1 may be omitted in the calculation. By deIinition, LE1
substances have no Ilammable cloud beyond the liquid pool. Eor toxic gases and toxic liquids all
relevant releases lead to a toxic exposure.
Outflow Release Type of Direct Delayed Event
frequency relevant release ignition ignition
(> 100 kg)
0.8 BLEVE
0.35
Instantaneous 1.0 Elash Iire
0.2
0.3 0.0 Explosion
0.8 Jet Iire
0.65
Table 3.1 Continuous 2" hole 1.0 Elash Iire
0.2
0.0 Explosion
0.7 No eIIect
Figure 3.1 Event tree pressurised flammable gas
Detailed QRA 3.4
Outflow frequency Relevant release Direct Delayed Event
(> 100 kg) (quantity) Ignition ignition
0.13 Pool Iire
0.15
Total inventory 0.0 Elash Iire
0.87
1.0 No eIIect
0.13 Pool Iire
Table 3.1 0.60
5.0 m
3
0.0 Elash Iire
0.87
1.0 No eIIect
0.25 No eIIect
0.5 m
3
Figure 3.2 Event tree atmospheric, flammable liquid LF2
Notes:
1. In a Iire, unburned toxic substances and toxic combustion products can be released to the
environment. In the case oI open Iires (open road situations), plume rise, due to the high
temperature oI the cloud, is assumed to occur immediately and no lethal eIIects are expected.
Exposure to unburned toxic substances and toxic combustion products do not have to be
considered in the QRA.
2. The release oI the internal energy oI a ruptured, pressurised transport unit can give rise to
blast waves and high velocity vessel Iragments. These eIIects do not have to be considered in
the QRA Ior an open road situation.
3. Substances that are both toxic and Ilammable should in principle be modelled using toxic
properties as long as the cloud is not ignited, and with Ilammable properties as soon as the
cloud ignites. However, this approach is currently too complicated Ior the models used. The
LOC is thereIore split into two independent events, namely, a pure toxic and a pure
Ilammable event (see section 4.7.3).
4. Eollowing the delayed ignition oI a vapour cloud, depending on obstruction, either a Ilash Iire
or blast overpressures (explosion) result. In open road situations the vapour cloud is
unconIined. In the QRA only the occurrence oI Ilash Iires need to be included, although in
densely populated or conIined areas the occurrence oI an explosion cannot totally be
excluded.
3.2.3 Modelling source term, dispersion, exposure and damage
AIter deIining the loss oI containment events and the resulting hazardous events, the source term
and the dispersion in the environment will have to be calculated, and the exposure and damage
determined. As the modelling in most cases does not diIIer Irom that described Ior stationary
Detailed QRA 3.5
installations at an establishment, only the speciIic aspects and diIIerences are highlighted in this
chapter.
Notes:
1. The continuous release Irom a pressurised transport unit is modelled as a hole in the vessel
wall with a sharp oriIice. The value oI the discharge coeIIicient should be set at Cd0.62.
Two-phase outIlow should be assumed. The release duration is to be determined using the
transport unit inventory and the mass Ilow rate. In the QRA calculation, the release duration
is limited to a maximum oI 30 minutes
2. The direction oI the continuous outIlow Irom a pressurised transport unit is set as horizontal,
parallel to the wind, because in an accident situation no speciIic inIormation is available. In
the case oI direct ignition oI a Ilammable gas, assume that the jet is unobstructed (Iree jet).
In all other cases (delayed ignition Ilammable gas, continuous outIlow oI liqueIied,
compressed toxic gases) obstructed outIlow should be assumed. The obstructed outIlow
should be modelled as a jet with the impulse reduced by a Iactor oI 4 and with dimensions
determined by the outIlow conditions.
3. Eor an instantaneous release oI liqueIied compressed gases, the mass in the vapour cloud
depends on the adiabatic Ilash Iraction, y, and is given by the relations in Table 3.3.
1able 3.3 Mass in vapour cloud following an instantaneous release
adiabatic Ilash Iraction, Mass in vapour cloud
(Iraction oI the total inventory oI the
transport unit)
0.1 2 *
0.1 0.36 * -0.028) / 0.26
0.36 1
4. lt is assumed that the direct ignition oI an instantaneous release oI liqueIied, compressed
Ilammable gas will result in a BLEVE. It has to be assumed that the total inventory oI the
transport unit takes part in the BLEVE.
5. The dimensions oI the pool Iormed, Iollowing a release oI liquid Irom an atmospheric
transport unit, are determined by a number Iactors: the discharge rate, the total volume oI
the release, the roughness oI the road and soil, the inclination oI the surIace and the possible
discharge into a sewer. As these Iactors are in most cases unknown, and no inIormation is
available, a Iixed pool may be assumed in the QRA. lt is recommended Ior a release oI the
total inventory oI a transport unit to assume a pool size oI 1200 m
2
, and Ior a release oI 5 m
3
to assume a pool size oI 300 m
2
|VeVoWeg96|.
6. The meteorological data to be used in a QRA is described in section 4.10. Data Irom a
representative meteorological station Ior the transport route under consideration should be
used. II necessary, data Irom diIIerent stations should be used Ior diIIerent road sections.
Detailed QRA 3.6
7. A representative value Ior the aerodynamic roughness length z0 Ior the terrain surrounding
the transport route, as described in section 4.6.2, should be used. II necessary, diIIerent
values should be used Ior diIIerent road sections.
8. Chemical processes, and the wet and dry deposition processes, do not have to be considered
in the modelling oI the vapour cloud dispersion.
9. Exposure and damage should be modelled as described in chapter 5 oI |RIVM99|.
Eor BLEVEs, a maximum exposure time oI 20 seconds in built-up areas as well is
recommended.
10. The population present in the surroundings oI a transport route should be surveyed according
to the rules outlined in section 5.3. The survey oI the population present within 300 meters
should be more detailed than at larger distances Irom the transport route. The risks along a
transport route are dominated by the transport oI Ilammable liquids and gases and the
eIIective distances oI the signiIicant scenarios are less than or comparable to 300 metres.
3.2.4 Calculation and presentation of results
The results oI a QRA are the Individual Risk and the Societal Risk and have to be presented
clearly as described in chapter 6 oI |RIVM99|. A widely used method Ior the calculation oI
risks is described in section 6. According to current regulations, the Societal Risk has to be
calculated and presented per kilometre oI transport route.
Notes:
1. The size oI the calculation grid should be small enough not to inIluence the calculation
results. The size oI the grid cell should not be larger than 25 x 25 metres, as the risks along
a transport route are dominated by the transport oI Ilammable liquids and gases and the
eIIective distances oI the signiIicant scenarios are less than or comparable to 300 metres. At
distances larger than 300 metres, a larger grid cell may be used. Consequently, the survey oI
the population present within 300 metres should be more detailed than at larger distances
Irom the transport route. It should be ensured that the distance oI the population Irom the
transport route is correctly represented.
2. Accident locations have to be selected at regular distances, and beyond the beginning and
end oI the road section under consideration, to produce a smooth risk contour. There should
be enough locations to ensure that the risk contour does not change substantially when the
number oI accident locations is increased.
3. In the calculation oI the Societal Risk, the choice oI the beginning and end oI the sections
Ior which the results are presented may be arbitrarily speciIied beIorehand. However,
sections should be chosen in such a way that the Societal Risk is maximised.
4. As well as the presentation oI the Individual Risk and the Societal Risk, other results can be
presented which will give a picture oI the risks. UseIul presentations include graphs oI the
Detailed QRA 3.7
probability oI death along the centre line as a Iunction oI distance Ior the dominant LOCs.
Also, those road sections with the highest risks or the road sections and locations which do
not comply with the risk criteria may be highlighted. The contribution oI the diIIerent
substances or substance categories, or the contribution oI the transport during day or night to
the overall risks may be speciIied.
3.2.5 Obstacles
A speciIic road section may be (partly) covered or a tunnel may be present. Also, obstacles such
as (noise) barriers may be present. SpeciIic road sections may be elevated (viaduct, bridge) or
located below ground level. The presence oI these obstacles and their inIluence on the risks
should, in principle at least, be qualitatively incorporated in the QRA.
These obstacles, depending oI the type oI obstacle and the hazardous substance involved,
inIluence the calculation oI the risks in several ways:
The initial accident Irequencies may diIIer Irom those on open road sections where obstacles
are not present. The probability oI a BLEVE in a tunnel may be larger than in open road
situations because oI the higher temperatures in case oI a Iire occurring in a tunnel.
The probability oI a large release in tunnels, and the dimensions oI the liquid pool Iormed,
diIIer Irom those in open road situations or in cases where other barriers are present or the
road is elevated.
The evaporation Irom liquid pools, and the dispersion oI the hazardous vapour cloud Iormed,
are inIluenced by the presence oI barriers. The ventilation and presence oI vehicles in
tunnels inIluences the evaporation.
The release oI toxic combustion products and heat resulting Irom Iires in tunnels and
enclosures must be considered.
The Iormation oI the vapour cloud Iollowing a release and Ilash oI liqueIied, compressed
gas is restricted by the enclosure. Heat transIer Irom tunnel and road surIaces however may
enhance the evaporation.
In enclosed situations the occurrence oI conIined explosion is possible. Also in tunnels, due
to the release oI internal energy, blast waves and pressure loads may result which lead to
Iailure oI the enclosure. Depending on the hazardous substance, the release scenario
involved, and the strength and openness oI the enclosure (tunnel, covering) these explosions
may or may not lead to the Iailure oI the enclosure, the Iormation and dispersal oI
Iragments. Consequently hazardous eIIects may occur at the ends oI the enclosure or along
the enclosure.
Due to the presence oI obstacles the heat radiation to the surroundings may be attenuated.
The inIluence on the risks oI the aspects mentioned above should be considered on a case by
case basis. No general rules can be given. There is yet no general consensus on how these
eIIects should be incorporated into a QRA. A preliminary model has been developed in the
Netherlands in the project ABIETO-Aanzet tot een Berekeningsmethodiek voor In- en Extern
risico bij Tunnels/Overkappingen (Initiative Ior a calculation method Ior internal and external
risk at tunnels/covered areas). In the commonly used integrated soItware packages the aspects
mentioned above can only be approximately addressed. Calculations with complex computer
codes (CED calculations) or experimental results are needed Ior a more reliable quantitative
determination.
Detailed QRA 3.8
3.3 Rail transport
3.3.1 Introduction
This chapter gives an outline oI a detailed QRA Ior the rail transportation oI dangerous
substances in bulk, in shunting yards, and in private sidings. The loss oI containment events
(LOCs) that need to be included in the QRA are identiIied. The corresponding Iailure
Irequencies and the various events Iollowing a release, and their probabilities, are given. The
modelling oI the source term, dispersion, exposure and damage and the calculation and
presentation oI the results is not described in detail. Only the diIIerences in modelling as
compared with stationary installations are highlighted.
The QRA rules are general and to be applied to railways, shunting yards and private sidings.
The basic data Ior a QRA are:
description oI the transport streams (number oI yearly loaded tank cars per substance or
category, during daytime and night-time);
description oI the transport units (characteristic inventory);
description oI the transport route;
description oI the shunting yard or the private siding;
description oI the activities on the shunting yard or the private siding;
accident Irequencies;
description oI the ignition sources;
properties oI transported (example) substances;
terrain classiIication oI the surroundings oI the transport route;
meteorological data;
population distribution in the surroundings oI the transportation route, the shunting yard or
the private siding.
are given in paragraph 4.4.
The risk oI the rail transportation oI dangerous substances is determined by the bulk transport oI
Ilammable gas (A), toxic gas (B2 en B3), toxic liquid (D3 and D4) and Ilammable liquid (C3).
All other substances are not relevant and do not have to be considered.
The LOCs Ior rail transportation are deIined |SAVE95, SAVE95a|. The LOC`s are
leakage Irom a 3" hole in the tank car;
rupture oI the tank car.

In this chapter a distinction will be made between accidents on a railway (outside shunting yards
and private sidings) and at shunting yards or private sidings. Although the LOCs are identical
Ior both situations, the appropriate LOC Irequencies are diIIerent. Eurthermore, distinction has
to be made between atmospheric and pressurised rail tank cars.
Detailed QRA 3.9
Notes:
1. LOCs Ior explosives and radioactive material will diIIer Irom the LOCs Ior other dangerous
substances. These transported substances are not normally included in the QRA as the
transport Irequency is low and there will thereIore be no measurable inIluence on the risk
level.
2. LOCs Ior small packages like drums and cylinders will be diIIerent to LOCs Ior bulk
transport. These transport movements are excluded Irom the QRA iI bulk transport takes
place. This is because oI their (too) little contribution to risk.
Eor the deIinition oI accident scenarios used to determine the accident Irequency, distinction has
to be made between private siding, shunting yards and railways.
3.3.3.1 Private siding
A private siding is a branch oI the railway Ior the delivery oI goods at a kind oI terminal. Its
characteristics are:
the sets oI points are manually operated;
the maximum speed is 30 km/hr;
level crossings with roads;
not equipped with signs.

The accident Irequencies have been determined |SAVE95a|. A distinction is made between the
shunting yard part oI the private siding and the railway part oI the private siding. The
Irequencies are:
1.1 * 10
-6
per car per km Ior the railway part;
8.4 * 10
-7
per car Ior the shunting yard part.
Eor the latter, the accident locations are spread over the shunting yard part.
The outIlow probability Ior atmospheric and Ior pressurised tank cars are given in Table 3.4:
1able 3.4 Outflow probability for atmospheric and pressurised tank cars
Scenario atm. tank car pressurised tank car
Collision 0.1 0.01
One car accident 0.1 0
3.3.3.2 Shunting yards
Eor shunting yards eight diIIerent scenarios are identiIied. Whether or not a scenario can take
place depends on the shunting process. The Iull list oI accident scenarios is:
sc. 1: a train collision upon arrival at or departure Irom a shunting yard;
sc. 2: a collision between an arriving/departing train and a set oI wagons;
Detailed QRA 3.10
sc. 3: a collision between a set oI wagons and a train that is shunted or Iormed;
sc. 4: a collision while changing locomotives;
sc. 5: an accident involving one tank car only;
sc. 6: a shunting accident;
sc. 7: intrinsic Iailure oI a tank car;
sc. 8: a BLEVE oI a tank car due to a leaking Ilammable liquid car nearby.
The accident Irequencies are given in table 3.5 Ior atmospheric and pressurised tank cars. The
probability oI an outIlow as a consequence oI the accident is added to this table. This
probability diIIers per type oI tank car and per scenario.
1able3.5 Accident frequency and outflow probability for a shunting yard
Scenario Accident frequency Probability outflow
atm. tank car Press. tank car
1. Arrival/departure
. ATP` present 5.5 * 10
-7
per train 0.1 0.01
. no ATP` present 5.5 * 10
-6
per train 0.1 0.01
2. Collision with set oI wagons 2.12 * 10
-5
per train 0.1 0.01
3. Shunting/Iorming 2.12 * 10
-5
per train 0.1 0.01
4. Changing locomotives 1.0 * 10
-6
per train 0.05 0.005
5. One car accident 2.75 * 10
-5
per train 0.1 0
6. Shunting accident 1.76 * 10
-5
per car 0.1 0.01
7. Intrinsic Iailure 5.0 * 10
-7
per car.year 1 1
8. BLEVE Eormula (note 4) 0 1
ATP: Automatic Train Protection
Notes:
1. The Irequency oI scenario 7 has additionally to be multiplied with the Iraction oI time that
the car is present (per year).
2. The BLEVE Irequency I on a shunting yard is calculated according to:
I Io . N . n . (Ap/A
tot
) . (t/12) . (T/365) . R
with
I : BLEVE Irequency per year
Detailed QRA 3.11
Io : Irequency oI a major Iire ( 3.1 * 10
-7
per Ilammable liquid car)
N : number oI Ilammable liquid cars per year.
n : number oI pressurised tank cars per train.
Ap : area oI Ilammable liquid pool ( 600 m
2
)
A
tot
: area oI shunting yard (m
2
)
t : hours that a pressurised car is present per day (is divided by 12 Ior the hours/day.
II day plus night is used, then the denominator should be 24.)
T : number oI trains with pressurised cars per year (is divided by 365 which is the
number oI days in a year)
R : repression Iactor ( 0.1)
This Iormula is based on the assumption that pressurised and atmospheric cars are present at the
same time. II not, the BLEVE Irequency equals zero.
Multiplication oI the accident Irequency with the outIlow probability will give the outIlow
Irequency. As only 10 oI the outIlows are relevant Ior external risks the outIlow Irequency has
to be multiplied with 0.1. This does not apply to scenarios 7 and 8 as they represent
instantaneous sources.
This approach results in outIlow Irequencies per train or per car. These Irequencies have to be
multiplied by the actual number oI trains and cars that are handled per year in the shunting yard.
lt was agreed |SAVE97| not to use the actual number oI trains but to put these numbers into
categories. The translation is carried out as presented in table 3.6.
1able 3.6 Categorisation of frequency per year of cars or trains
Actual Irequency range Erequency per
per year year Ior the
calculation
1 - 12 0
13 - 50 50
51 - 100 100
101 - 200 200
201 - 400 400
401 - 800 800
801 - 1600 1600
etc. etc.
When trains consist only oI cars with the same dangerous substance, no Iurther correction is
necessary.
II the train has a mixed composition and the Irequency is given per train, then the LOC
Irequency has to be multiplied by the Iraction oI cars in the train with the dangerous substance
under consideration.
Detailed QRA 3.12
An outIlow (larger than 100 kg) may lead to a continuous source or an instantaneous source
(excluding scenarios 7 and 8). The probability oI a continuous source is 0.6 and that oI an
instantaneous source is 0.4.
In cases oI Ilammable gas or liquid, an immediate ignition is possible. This probability is set at
0.8.
The locations oI possible accidents are equally spread over that part oI the shunting yard where
the activity under consideration actually takes place.
3.3.3.3 Railways
The generic accident Irequency Ior a rail car on railways outside the shunting yard is 3.6 * 10
-8
per car kilometre |SAVE95|. II it is known that the train speed is more than 40 km/hr, then this
Irequency has to be corrected by multiplying by 1.26. II the speed is lower than 40 km/hr then
the correction Iactor is 0.62.
II there is a set oI points on the railway, the accident Irequency is increased by 3.3 * 10
-8
per car
km.
Eor a level crossing the addition is 0. 8 * 10
-8
per car km.
So:
Eor 1 km track with 1 set oI points : 6.9 10
-8
per car
Eor 1 km track with 1 level crossing : 4.4 10
-8
per car.
Eor dedicated chlorine trains in the Netherlands the accident Irequency is multiplied by 0.2
because oI extensive saIety measures.
Given an accident occurs, probabilities Ior a release oI more than 100 kg have been set. A
distinction is made between sections where train speeds are allowed to be more than 40 km/hr
and where they must be less than 40 km/hr.
Table 3.7 presents the results.
1able 3.7 Probability of outflow (~ 100 kg) given an accident
Allowed Probability oI outIlow (> 100 kg)
Train speed Atmospheric tank cars Pressurised tank cars
40 km/hr 7.9 * 10
-2
7.9 * 10
-4
> 40 km/hr 5.6 * 10
-1
2.8 * 10
-3
Given a relevant outIlow, the probability oI an instantaneous source is 0.4 and oI a continuous
source is 0.6.
Detailed QRA 3.13
Eor Ilammable gases the probability oI direct ignition is set at 0.8 Ior an instantaneous source
and 0.5 Ior a continuous source. Eor Ilammable liquids the probability is 0.5.
II there are cars with Ilammable liquid in a train oI mixed composition, the BLEVE risk Ior cars
with Ilammable or toxic gas increases. The reason Ior this is the possibility oI a burning pool
under the car. This risk increase is incorporated as Iollows:
. iI a train consists oI cars with Ilammable gas and oI cars with Ilammable liquid then the
BLEVE Irequency is multiplied with (N l),
where N is the number oI Ilammable liquid cars per train;
. iI a train consists oI cars with toxic gas and oI cars with Ilammable liquids, the Irequency oI
an instantaneous release is multiplied by (0.8 N l), where N is the number oI Ilammable
liquid cars per train.
II a train consists only oI cars with a liqueIied Ilammable gas then the BLEVE Irequency is not
increased. |SAVE95| shows that the extra contribution is negligible.
AIter deIining the loss oI containment events and the resulting hazardous events, the source
term and the dispersion in the environment will have to be calculated and the exposure and
damage has to be determined. As the modelling in most cases does not diIIer Irom those
described Ior stationary installations at establishments (chapter 6 oI |RIVM99|), only the
speciIic aspects and diIIerences are highlighted in this chapter.
Notes:
1. Eor atmospheric tank cars, outIlow is not calculated. lt is agreed that a continuous source
results in a pool oI 300 m
2
and an instantaneous source in a pool oI 600 m
2
. The source term
Ior dispersion is equal to the calculated evaporation rate. The assumptions Ior the evaporation
calculations are that the depth oI the gravel layer is 0.15 m and the gravel diameter is 1 CED.
2. Eor pressurised liqueIied gases the double Ilash percentage gives the vapour source in cases
oI an instantaneous source. The Iluid, not Ilashed, will evaporate according to the description
given in note 1. The amount that evaporates during one minute should be added to the source
term. Eor leakage, Iull evaporation is assumed.
3. The inventory oI pressurised tank car is set at:
- Ilammable gas : 48 tonnes
- toxic gas : 50 tonnes
Eor the Iuture (2010) larger inventories are expected. The Ilammable gas tank car may
contain 54 tonnes, chlorine tank cars 60 tonnes and ammonia tank cars 58 tonnes.
4. A direct ignition oI an instantaneous release oI Ilammable gas results in a BLEVE oI the
total inventory.
5. A direct ignition oI a continuous release oI Ilammable gas results in a jet Iire comprising the
total outIlow rate.
6. The continuous outIlow is calculated as a Ilow out oI a 3" hole with a contraction coeIIicient
oI 0.67.
representative meteorological station Ior the transport route or the shunting yard under
consideration should be used. II necessary, data Irom diIIerent stations should be used.
Detailed QRA 3.14
8. A representative value Ior the aerodynamic roughness length Zo Ior the terrain surrounding
the transport route or the shunting yard as described in section 4.6.2 should be used. II
necessary, diIIerent values should be used Ior diIIerent sections. DeIault value z
o
1.0 metre.
9. Chemical processes and the wet and dry deposition processes do not have to be considered
10. Exposure and damage should be modelled as described in chapter 5.
11. The population present in the surroundings oI a transport route or the shunting yard should
be surveyed according to the rules outlined in section 5.3.
The results oI a QRA are the Individual Risk Contours and the Societal Risk Graphs. A widely
used method Ior the calculation and presentation oI risks is described in chapter 6 oI |RIVM99|.
According to current regulations the Societal Risk has to be calculated and presented per
kilometre oI transport route. Eor shunting yards this does not oI course apply.
Notes:
results. The size oI the grid cell should not be larger than 25 metres as the risks along a
transport route are dominated by the transport oI Ilammable liquids and gases and the
eIIective distances oI the signiIicant scenarios are less than or comparable to 300 metres.
At distances greater than 300 metres, a larger grid cell may be used.
2. Accident locations on the transport route have to be selected at regular distances and beyond
the beginning and end oI the section under consideration to produce a smooth risk contour.
There should be enough locations to ensure that the risk contour does not change
substantially when the number oI accident locations is increased.
3. In the calculation oI the Societal Risk Ior transport routes the choice oI the beginning and
end oI the sections Ior which the results are presented may be arbitrarily set beIorehand.
However, sections should be chosen in such a way that the Societal Risk is maximised.
3.4 Inland waterway transport
3.4.1 Introduction
This chapter describes the QRA Ior the transport oI dangerous substances in bulk by inland
waterways. The QRA rules given are general and applicable to main inland waterways oI normal
nautical complexity. Situations in which speciIic nautical circumstances occur need speciIic
consideration, such as passing through locks or at harbours.
The basic data needed in conducting a QRA Ior a speciIic inland waterway include the:
Description oI the inland waterway (navigability class, navigable width and overall width)
Description oI the transport streams (annual Irequency oI loaded transport units per
substance)
Description oI the number oI accidents and amount oI traIIic in order to determine accident
Irequencies
Description oI the transport units (characteristic inventory)

Detailed QRA 3.15
Properties oI transported substances
Terrain classiIication oI the surroundings oI the waterway
Meteorological data
Population present in the surroundings oI the waterway

or previous chapters are given in this chapter and chapter 4.
The risks oI the transport oI dangerous substances by inland waterways are determined by
numerous Iactors, the most important oI which are:
type oI containment oI the substance e.g. type oI ship or tank
substance properties e.g. chemical, physical, toxicological
waterway characteristics e.g. traIIic density and composition, curvature, navigable width,
current
As Ior all transport units the LOCs Ior external impact e.g. collisions with other ships or
objects, dominate the risks oI moving vessels. In a harbour this may oI course be diIIerent. A
check on the contribution oI other LOCs, such as intrinsic Iailure, may then be appropriate.
Type oI containment
QRAs oI inland waterways are limited to the transport oI dangerous liquids and gases in bulk.
Solids are not considered due to the minor eIIect they have on the external risk. Transport in
containers is not considered because oI the limited quantity oI a packing unit and the reduced
probability oI outIlow compared to other vessels.
The Iollowing vessel types are considered:
Single hull vessel
Double hull vessel (transport at room temperature or reIrigerated transport)
Gas tanker
Typical loading capacities oI vessels range Irom 300 up to 2000 tonnes. Individual tanks may
contain up to 200 m
3
oI inventory.
Properties oI substances
QRAs oI inland waterways are limited to the transport oI dangerous liquids and gases. Solids
are not considered though they may present an environmental problem or a problem oI
occupational saIety. The transport oI explosives and radioactive materials on inland waterways is
extremely rare and is hitherto not considered in a QRA.
Waterway characteristics
Waterway characteristics are a part oI the QRA. The level oI detail depends on the required
thoroughness oI the analysis. At least the Iollowing needs to be known:
overall width oI the waterway
navigable width oI the waterway
the distribution oI the traIIic on it

The LOCs Ior inland waterway transport are Irom |V&W89, SAVE88, AEA95| and given below.
Detailed QRA 3.16
1able 3.8 LOCs for external impact (traffic incidents) of moving vessels
Vessel type Outflow Volume Duration (min)
Single hull continuous minor 30 m
3
Irom 150 m
3
tank 30
continuous major 75 m
3
Irom 150 m
3
tank 30
Double hull or reIrigerated continuous minor 20 m
3
Irom 150 m
3
tank 30
continuous major 75 m
3
Irom 150 m
3
tank 30
Gas tanker
1
) continuous minor hole diameter 3" in 180 m
3
30
tank
continuous major hole diameter 6" in 180 m
3
30
tank
1
) Temperature is an important parameter in calculating the outIlow rate. Pressurised transport at a decreased
temperature (but not totally reIrigerated) is sometimes indicated as 'semi-gas. The same LOCs are applicable
(see also note 1 below).
Notes:
1. The scenario considered is the rupture oI a pipe connected to the tank Iollowing a dislocation
oI the tank due to the impact. The largest pipe diameter connected to the liquid phase is 6".
OutIlow should be calculated as two phase Ilow. The location oI the rupture is conservatively
taken as close to the pressure vessel wall. The release is modelled as a hole with a sharp
oriIice (C
d
0.62). The direction oI outIlow is set as horizontal, parallel to the wind. The
temperature (and saturation pressure) oI the transported medium should be taken into
account. Eor example, Ior the transport oI ammonia it is oIten taken to be 5C.
Sometimes this state is reIerred to as a separate vessel type called 'semi-pressurised or
'semi-gas tanker.
2. The typology oI vessels is relatively approximate. Since the introduction oI the LOCs in the
late eighties, ship-technology and classiIication demands have developed. The ADNR
|ADNR|, the international treaty on the transport oI dangerous substances in the river Rhine
area, lays down the construction demands on the Iollowing vessel types:
N-vessels (Rn. 331.000)
C-vessels (Rn. 321.000)
G-vessels (Rn. 311.000)
Dry bulk single hull vessels (Rn 110.000)
Dry bulk double hull vessels (Rn. 110.288)

The last two vessel types are not considered in the QRA. G-vessels are gas tankers, C-vessels
are double hull vessels. N-vessels are considered as 'single hull vessels. Some conservatism
is built in here as the category N-vessels contains pure single hull vessels as well as single
hull vessels with independent tanks.
The deIault link between substance categories and ship types is given in Table 3.9.
Detailed QRA 3.17
1able 3.9 Relation between substance category and ship type
Substance category Ship type
LE1, LE2 Single hull
LT1, LT2 Double hull
GE, GT Gas tanker
The minimum allowable tank ship type Ior a speciIic substance is laid down in addendum 4
oI appendix B2 oI the ADNR.
II it happens that substances are transported more saIely than the minimum, e.g. a C-type
vessel is used where a N-type is in the list, this may be taken into account in a detailed
QRA. Eor substance categories where N- as well as C-vessels are possible, e.g. the
categories oI the Ilammable liquids, the list oI substances oI the ADNR may be
conservatively used Ior an estimation oI the ratio single hull/double hull.
3. The LOCs oI Table 3.8 are applicable in cases where the Iraction oI seagoing vessels is
below 10. Eor collisions between seagoing vessels, and between seagoing vessels and
inland vessels, diIIerent LOCs should be considered |PROT98|.
3.4.3 Accident frequency and outflow probabilities
The Irequency oI a hazardous event at a speciIic location on the transport route per transport unit
per year can be represented as the product oI: an initial accident Irequency (per unit km), the
probability oI signiIicant release given an accident Ior the speciIic transport unit, and the
probability oI the hazardous event given a release.
Initial accident Irequency
The initial accident Irequency is the Irequency oI serious damage to a vessel per unit distance,
(vessel kilometre). The basic data needed to estimate the accident Irequency are described in
chapter 4.5. Typical Ior N-vessels are values between 10
-4
and 10
-7
per vessel kilometre per year
|AVIV93|.
The initial accident Irequency is strongly dependent on local Iactors. A generic approach to the
QRA should thereIore be conducted with care.
In Table 3.10 Irom |IPORBM|, deIault values Ior the initial accident probability are given. As
stated beIore, these Irequencies should be used with the utmost care. Whenever possible
Irequencies should be derived Irom local data. The method is described in chapter 4.5.
Local accident Irequencies Ior a number oI locations along main waterways in the Netherlands
may be Iound in |AVIV93|. The navigability class Ior the main inland waterways in the
Netherlands is given in Table 3.11.
Detailed QRA 3.18
1able 3.10 Default values for the initial accident frequency as a function of navigability class
Navigability class (CEMT) Initial accident frequency (/vessel km)
4 6.7*10
-7
5 7.5*10
-7
6 1.4*10
-6
1able 3.11 Navigability-class of some Dutch inland waterways
Waterway CEMT class Waterway CEMT class
Eemskanaal 5 Waal 6
v Starckenborghkanaal 5 Beneden Merwede 6
Prinses Margrietkanaal 5 Noord 6
IJssel 5 Hollandsch Diep 6
Nederrijn 5 Schelde-Rijnkanaal 6
Lek 5 Maas 5
Amsterdam-Rijnkanaal 6 Julianakanaal 5
Nieuwe Maas 6 Kanaal Gent-Terneuzen 6
Oude Maas 6 Hollandsche IJssel 5
Oude Rijn/Gouwe 4
Probability oI release
The probability oI release given serious damage (or, Ior G-vessels, more severe damage) is
taken Irom |V&W89| and given in Table 3.12.
1able 3.12 Probability of release given serious damage
Vessel type Outflow Probability given serious damage
Single hull Continuous minor 0,2
Continuous major 0,1
Double hull or reIrigerated Continuous minor 0,006
Continuous major 0,0015
Gas tanker (irrespective oI, Continuous minor 0.025
transported substance
temperature)
Continuous major 0.00012
Probability oI hazardous event
The events to be considered in the QRA are
Toxic exposure
Pool Iire
Detailed QRA 3.19
Jet Iire
Elash Iire
Eor the QRA, deIault immediate ignition probabilities are used as given in Table 3.13. In
Societal Risk calculations delayed ignition should be modelled as a Iunction oI the distribution
oI ignition sources (detailed QRA Ior existing/planned situation). When the speciIic locations oI
ignition sources are unknown the values Irom Table 3.13 may be used instead (QRA Ior a
generic situation). Delayed ignition in a Individual Risk calculation should be modelled to give
the maximum eIIects. Ignition at maximum cloud area should be assumed.
1able 3.13 Ignition probability
Substance Ignition probability
Immediate Delayed
Elammable liquid, category LE2 0.065 0.065
Elammable liquid, category LE1 0.01 -
Elammable Gases 0.5 0.1
Notes:
1. As already stated, initial Irequencies may vary widely due to local Iactors. The question is
how to decide whether a speciIic location 'deserves its own accident Irequency. Two Iactors
are relevant to making this decision:
Nautical judgement;
Historical data.
Review oI the distribution oI locations oI accidents along the waterway may indicate possible
concentration points. A nautical expert judges whether (and which) speciIic local Iactors are
present that may increase the accident Irequency. II so, the accidents (or some oI them) are
considered speciIic Ior that location. II not ('the accidents might as well have happened
anywhere else along the waterway) the accidents are considered representative Ior the total
length oI waterway being studied. More detail on the retrieval oI accident Irequencies Irom
accident data is given in chapter 4.5.
2. Nautical judgement is necessary in evaluating the eIIect oI local measures on the accident
Irequency. Local measures are, Ior example, restrictions in overtaking, moving speed,
anchoring. A traIIic and accident model has been developed |MSCN95| to assist in
quantiIying the eIIects oI these kind oI measures. It considers the eIIect oI relevant Iactors
like traIIic composition, navigable width, quay-related activities etc. on the accident
Irequency. Use oI the model, however, requires some nautical knowledge. The magnitude oI
the eIIect oI measures on the accident Irequency is limited to about 20 |WL95|.
AIter deIining the loss oI containment events and the resulting hazardous events, the source
term and the dispersion in the environment will have to be calculated, and the exposure and
damage determined. As the modelling in most cases does not diIIer Irom that described Ior
stationary installations at establishments in chapters 4 and 5 oI |RIVM99|, only the speciIic
aspects and diIIerences are highlighted in this chapter.
Detailed QRA 3.20
Notes:
1. Two-phase outIlow should be assumed. The release duration is to be determined using the
transport unit inventory and the mass Ilow rate. In the QRA calculation, the release duration
is limited to a maximum oI 30 minutes.
2. The pool dimensions are determined by the equilibrium between the discharge rate, the
evaporation rate, the solution rate and the rate oI spreading under gravity. Models are
described in |SAVE88|.
The presence oI current elongates the pool. The maximum length is u*t
max
, where u is
velocity (m/s) and t
max
is the time (s) when discharge rate and the loss terms reach
equilibrium. In case oI (immediate) ignition the equilibrium between burning rate and
discharge rate determines the pool diameter as implemented in |IPORBM|.
3. Other physical and chemical processes:
Sinking liquids are not considered in the QRA.
Evaporation models oI liqueIied gases |SAVE88|, neutral or buoyant in water, are seldom
used. Purely liqueIied transport at atmospheric pressure does not actually take place.
Some substances react with water to produce Ilammable or toxic gases. Models to predict
the release oI Ilammable (mostly hydrogen) and or toxic gases (mostly hydrochloric acid)
oI substances that react with water have not been described. These substances (mostly
substances oI class 4.3 and 8 oI the AD(N)R) are seldom transported in bulk.
4. Vapour cloud dispersion
The direction oI outIlow is set as horizontal, parallel to the wind, because in an accident
situation no speciIic inIormation is available.
In the case oI direct ignition oI a Ilammable gas, assume that the jet is unobstructed (Iree
jet).
In all other cases (delayed ignition Ilammable gas, continuous outIlow oI liqueIied,
compressed toxic gases) obstructed outIlow should be assumed. The obstructed outIlow is
modelled as a jet with the impulse reduced by a Iactor 4 and with dimensions determined by
the outIlow conditions.
6. Exposure and damage should be modelled as described in chapter 5 oI |RIVM99|.
used. II necessary, data Irom diIIerent stations should be used Ior diIIerent waterway
sections.
8. A representative value Ior the aerodynamic roughness length Z
0
Ior the terrain surrounding
the transport route as described in section 4.6.2 should be used. II necessary, diIIerent values
should be used Ior diIIerent waterway sections.
Detailed QRA 3.21
9. Eollowing the delayed ignition oI a vapour cloud, depending on obstruction, either a Ilash
Iire or blast overpressures (explosion) result. Usually in waterway accidents the vapour
cloud is unconIined. In the QRA only the occurrence oI Ilash Iires needs to be considered.
10. The population present in the surroundings oI a transport route should be surveyed
according to the rules outlined in section 5.3. Toxic substances dominate the societal risk. In
cases where no toxic substances are being transported along a certain route, the population
survey may be restricted to 200 metres Irom the waterway rim, but should be rather detailed
e.g. a grid size oI 25 metres should be used. When toxic gases are present, the survey should
be extended to 2000 metres Irom the waterway rim. Eor distances over 200 metres a larger
grid size e.g. 100 metres is suIIicient. The same applies also beyond the beginning and the
end oI the calculation area.
3.4.5.1 Individual Risk
The results oI a QRA are the Individual Risk and the Societal Risk and have to be presented
clearly as described in section 6. A widely used method Ior the calculation oI risks is described
in section 6.
The spacing oI accident locations should be less than or equal to 25 metres because Ilammable
liquids dominate the risk at short distances. The resulting risk contours should be smooth and
should not reIlect discontinuities due to a too wide spacing oI accident locations.
The same applies to the edges oI the calculation area. Contours Ior limited sections that
intersect with the waterway should not be presented. The calculated area should thereIore be
larger than the area to be presented in the report. When toxic substances are present, usually
about 1000 metres extra on both sides is needed. In cases where only Ilammables are present,
usually about 200 metres extra on both sides is enough. The criterion is that the risk contour
does not change substantially when the number oI accident locations is increased.
Attention should be paid to the distribution oI the traIIic over the width oI the waterway. A
uniIorm distribution is applicable unless speciIic reasons Ior some other distribution are present,
such as a limitation in passage by bridges, a prohibition in passing or landing etc.
3.4.5.2 Societal Risk
Calculation oI societal risk should be perIormed as stated in chapter 6.2 oI |RIVM99|. The
remarks on the spacing oI accident locations and the distribution oI the traIIic on the waterway
apply here as well. Care should be taken to avoid overlap between populated grid cells and the
waterway (most programs check on this).
The size oI the area to be surveyed Ior numbers oI inhabitants depends on the composition oI
traIIic as given in Table 3.14, derived Irom |AEA95| and |IPORBM|.
Detailed QRA 3.22
1able 3.14 Distances from waterway-edge to survey population for societal risk calculation
Substances present Distance from edge (m)
Elammable liquids 50
Elammable gases 500
Toxic liquids 500
Toxic Gases 2000
Societal risk, when compared to the Dutch government standards, should be calculated and
presented per kilometre route. Decisive Ior the comparison is the kilometre that produces the
maximum societal risk.
3.4.6 Presentation of results
Results should be presented as stated in chapter 6.3 oI |RIVM99|. As well as the presentation oI
the Individual Risk and the Societal Risk, other results can be presented which will give a
picture oI the risks. UseIul presentations include graphs oI the probability oI death along the
centre line as a Iunction oI distance Ior the dominant LOCs. Also those waterway sections with
the highest risks or the waterway sections and locations which do not comply with the risk
standards may be highlighted. The contribution oI the diIIerent substances or substance
categories, or the contribution oI the transport during daytime or night-time to the overall risks
may be speciIied.
3.5 Pipeline transport
3.5.1 Introduction
This chapter gives an outline oI a QRA Ior the transport oI dangerous substances by pipeline.
The loss oI containment events (LOCs) that need to be included in the QRA are identiIied. The
corresponding Iailure Irequencies and the various events Iollowing a release and their
probabilities are given. The modelling oI the source term, dispersion, exposure and damage, and
the calculation and presentation oI the results is not described in detail. Only the diIIerences in
modelling as compared with stationary installations are highlighted.
The data required Ior a QRA are:
Description oI the transport system (diameter, location emergency valves);
Description oI the transport stream (substance, Ilow rate);
Description oI the ignition sources;
Properties oI transported substances;
Terrain classiIication oI the surroundings oI the transport route;
Meteorological data;
Population distribution in the surroundings oI the transportation route.

are given in paragraph 4.5.
Detailed QRA 3.23
Notes:
1. This chapter only applies to underground cross-country pipelines.
2. This chapter does not always apply to the pipeline transport oI methane and Ilammable
liquids (K1, K2 or K3). Eor these substances special tables apply that prescribe the distance
between a pipeline and nearby housing, iI deIined saIety measures are taken. These tables are
implemented in |IPORBM|.
3.5.2 Loss of Containment Events
Eor underground pipelines two LOCs are deIined |SAVE95b|. The LOCs are:
- leakage Irom a 20 mm hole in the pipeline;
- rupture oI the pipeline.
The accident Irequency equals the Irequency oI outIlow and is given per km per year. Table 3.15
presents the accident Irequency Ior the diIIerent types oI underground pipelines |SAVE95b|.
1able 3.15 LOC frequencies for various types of pipelines
Pipeline type Accident Erequency (1/km.yr)
Line located in a 'lane 7.0 * 10
-5
NEN 3650 - line 6.1 * 10
-4
All other lines 2.0 * 10
-3
Notes:
1. Line located in a 'lane means a pipeline located in a group oI pipelines, on a dedicated
route. LOC Irequencies Ior this situation are lower because oI extra preventive measures.
2. Eor speciIic pipelines other Irequencies may be used such as iI additional preventive
measures are proving to decrease the LOC probability. (This applies Ior instance to a chlorine
line in the Rijnmond area.)
The probability oI leakage or rupture, given an LOC, is given in table 3.16.
1able 3.16 Leakage and rupture probability, given an LOC
Pipeline type Leakage probability Rupture probability
Line located in a lane 0.9 0.1
All other pipelines 0.75 0.25
Detailed QRA 3.24
The hazardous events Iollowing an LOC are:
- toxic exposure
- jet Iire
- Iireball
- pool Iire
- Ilash Iire
- explosion.
The immediate ignition probabilities Ior LOC are given in table 3.17 |SAVE95b|.
1able 3.17 Immediate ignition probability for a LOC
Type oI substance Probability oI immediate ignition
Leakage Rupture
Elammable gas 0.04 0.09
LiqueIied Ilammable 0.14 0.30
gas
The probability oI a delayed ignition is, at the most, equal to: (1 - probability oI immediate
ignition), because there is a probability oI no ignition at all.
Note:
1. A higher probability oI direct ignition may be considered in the case oI hydrogen, given its
low minimal ignition energy.
The next step, Iollowing completion oI the LOCs, concerns the calculation oI the source term.
Eor this calculation a distinction has to be made between leakage and rupture.
3.5.4.1 Leakage
II there is no possibility that the LOC will be noticed, and actions to close (emergency) valves
can not take place or might Iail, the leakage is considered to be a continuous source lasting Ior
30 minutes.
II (emergency) valves are closed, then Ior
- gas: release will continue until the pressure in the pipeline equals the atmospheric pressure;
- liquid: release will take place due to liquid expansion until the vapour pressure in the pipeline
equals atmospheric pressure. II the pipeline is not level, outIlow due to gravity Iorces
should be considered;
- liqueIied gas: two phase outIlow occurs.
Detailed QRA 3.25
Note:
1. A comparison between the outIlow rate and the pump rate gives some inIormation about
whether the leakage might be noticed by the pipeline operator.
3.5.4.2 Rupture
In case oI a rupture, the calculated outIlow should be doubled because oI two-side outIlow.
Eurthermore, the outIlow should be increased by the pump rate until the pump is stopped or
valves are closed. Gravity Iorces (liquid lines) should also be taken into account.
Notes:
1. It should be checked whether the pump rate will increase because oI the pressure drop due
to the pipeline rupture.
2. Crater Iormation and location oI the hole (direction oI outIlow) is not taken into account
because oI lack oI reliable data.
3. The agreed maximum pool size is 3000 m
2
with a pool depth oI 0.1 m.
4. Given the critical temperature oI ethene, the outIlow oI ethene in the event oI a pipeline
rupture is sensitive to the assumed air and soil temperature |SAVE95b|. Eor leakage it is
shown that there is hardly a diIIerence in risk, whether the outIlow is calculated as a gas
Ilow or a two phase Ilow.
5. The evaporation Iactor Ior soil equals 1800 W/K.s
1/2
.
selected. II necessary, data Irom diIIerent stations should be used Ior diIIerent sections oI the
pipeline.
7. A representative value Ior the aerodynamic roughness length Z
0
Ior the terrain surrounding
the transport route as described in section 4.6.2 should be selected. II necessary, diIIerent
values should be used Ior diIIerent line sections. DeIault value Z
0
1.0 metre.
9. Exposure, and damage should be modelled as described in chapter 5.
10. The population present in the surroundings oI a transport route should be surveyed
according to the rules outlined in section 5.3.
AIter deIining the source term, the modelling does not diIIer Irom the modelling described Ior
stationary installations.
Detailed QRA 3.26
The results oI a QRA are the Individual Risk Contours and the Societal Risk Graphs. A widely
used method Ior the calculation and presentation oI risks is described in section 6. According to
current regulations, the Societal Risk Graphs have to be calculated and presented per kilometre
oI transport route.
Notes:
results.
2. Accident locations have to be selected at regular distances and beyond the beginning and end
oI the section under consideration to produce smooth risk contours. There should be enough
locations to ensure that the risk contour does not change substantially when the number oI
accident locations is increased.
3. In the calculation oI the Societal Risk, the choice oI the beginning and end oI the sections Ior
which the results are presented may be arbitrarily set beIorehand. However, sections should
be chosen in such a way that the Societal Risk is maximised.
Detailed QRA 3.27
4. BASIC DATA
4.1 Introduction
To perIorm a QRA oI a technical system the Iirst step is deIining this system, its characteristics,
and its boundaries. Transport systems present some special diIIiculties in this respect. There is a
wide variety oI possible accident-site characteristics and also a wide variety oI transport units
and substances may be involved.
The basic data needed in conducting a QRA Ior a speciIic transport route include the:
Description oI the transport route (location, type oI route, obstacles present)
Description oI the transport streams (annual number oI transport units per substance or
category, during daytime and night-time)
Description oI the number oI accidents and traIIic intensities in order to determine accident
Irequencies
Description oI the transport units (type oI unit, characteristic inventory)
Properties oI transported (representative) substances
Terrain classiIication oI the surroundings oI the transportation route
Meteorological data
Population present in the surroundings oI the transportation route

In chapter 3 methods Ior obtaining these data, recommended values, and reIerences to
inIormation sources Ior most oI the data items mentioned above were given. The method Ior the
categorisation oI substances used in IPORBM and in a QRA in the Netherlands is described in
paragraph 4.2. The application oI this method to a QRA Ior road and inland waterway transport
oI hazardous materials will be outlined. Eurthermore, the subsequent paragraphs will describe
how data on the transport oI hazardous substances in the Netherlands may be obtained, and how
a location speciIic accident Irequency may be deduced Ior a speciIic section oI a transport route
Irom the available databases in the Netherlands.
4.2 Categorising substances
4.2.1 Description of method and criteria
Due to the wide variety oI substances that are transported, there is a need to deIine a limited
number oI substance categories, and to classiIy each transported substance within this scheme.
Quantitative risk calculations may then be based on these categories and on a substance
representative oI a particular category. The method has been described in |AVIV95| and
|AVIV99|. This classiIication is based on the state oI aggregation, volatility, Ilammability and
toxicity. Eor transport on waterways, some additional properties are used i.e. solubility,
reactivity with water and relative density. The scheme is related to the classiIications Ior the
transport oI hazardous materials as applied in the ADR, RID and ADNR. The Iour main
categories are given in Table 4.1.
Basic Data 4.1
1able 4.1 Main substance categories for risk calculations
Category Description
GE Elammable gas
LE Elammable liquid
GT Toxic gas
LT Toxic liquid
In each main category, one or more subcategories are deIined and labelled numerically; a higher
number indicates a higher risk potential. Eor example, an LT4 substance is more dangerous`
than an LT1 substance. Substances that are both Ilammable and toxic are indicated by a
combined category, Ior example Ethylene oxide (UN 1040) is categorised as GE1/GT3.
The divisions and subdivisions oI the substance categories Ior a risk calculation are given below.
The subcategories GEO and GTO include compressed gases or reIrigerated liquids. These
categories, in most cases oI practical interest, can be omitted in a QRA.
Definitions
DeIinition state oI substances
Gas T
boil
293 K
Liquid T
boil
> 293 K and T
melt
293 K
Solid T
melt
> 293 K
DeIinition oI parameters
LC
50
LC
50
ihl rat 1 hr |ppm|
T
crit
critical temperature |K|
T
boil
boiling temperature at 1 bar |K|
P
20
vapour pressure at 20C |mbar|
T
Ilash
Ilash point |K|
T
melt
melting point |K|
Subdivisions
GE Elammable gas IMDG 'Ilammable
ADR/RID class 2.2, 2.4 ; item number E, TE, ETC
T
crit
293 K GE0 T
boil
182 K
T
crit
> 440 K GE1 T
boil
> 273 K
T
crit
400 440 K GE2 T
boil
253 273 K
T
crit
293 400 K GE3 T
boil
182 253 K
Basic Data 4.2
GT Toxic gas LC
50
5.10
4
T
crit
10
2
10
2
10
3
10
3
10
4
10
4
5.10
4
T
boil
T
crit
293 K GT0 GT0 GT0 GT0 T
boil
182 K
T
crit
> 440 K GT5 GT4 GT3 GT2 T
boil
> 273 K
T
crit
400 - 440 K GT5 GT5 GT4 GT3 T
boil
253 273 K
T
crit
293 - 400 K GT5 GT5 GT5 GT4 T
boil
182 253 K
LE Elammable liquid T
Ilash
334 K (61 C)
LE1 T
Ilash
> 296 K (>23 C)
LE2 T
Ilash
296 K (>23 C)
LT Toxic liquid LC
50
5.10
3
P
20
10
1
10
1
-10
2
10
2
-10
3
10
3
-5.10
3
T
boil
P
20
10 LT2 LT1 T
boil
> 373
P
20
10 50 LT3 LT2 LT1 T
boil
353373
P
20
50 200 LT4 LT3 LT2 LT1 T
boil
323353
P
20
200 700 LT5 LT4 LT3 LT2 T
boil
303323
P
20
> 700 LT6 LT5 LT4 LT3 T
boil
303
Additional points for transport on water
Eor transport on waterways, the Iollowing speciIic additional points should be addressed:
Sinking substances, that is substances with a relative density to water at 20C > 1.03, are
disregarded;
Substances reacting with water to yield toxic or Ilammable gases do not have to be
considered in the QRA. The amount oI substances transported (in bulk) is small. Also
speciIic source term models have not been developed or incorporated in commonly used
integrated soItware packages.
In case the relative density to water at 20C is less than 1,03 and P
20
is less than 700 mbar
the solubility oI the substance should be considered, Ior example in the source term and
eIIect calculation oI the QRA. Alternatively, as a good approximation, the substance category
number may be lowered by 1. Eor example, a readily dissolving substance (i.e. solubility >
4000 mol/m
3
) oI LT2 can be considered as LT1.
4.2.2 Use of method in a QRA
The most precise result in a QRA is oI course achieved by calculating Ior all substances.
However, this requires a large amount oI input data such as substance properties data. The
method may be applied at two diIIerent levels oI detail depending on the demands oI the project
involved |AVIV95, AVIV99|.
II a quick indication oI the risk levels is suIIicient, one, representative substance Ior each
substance category involved can be used to perIorm the calculation. The recommended
representative substances are summarised in Table 4.2.
Basic Data 4.3
OIten, a Iew speciIic substances account Ior the bulk oI transport in a substance category. A
more precise result may then be achieved by calculating these substances as such and
approximating only the last 20 oI the transported substances by using a representative
substance (80/20-rule).
1able 4.2 Representative substances per substance category
Substance category Representative substance
LE1 Nonane
LE2 n-Pentane
LT1 Acrylonitrile
LT2 Nitric acid (70) *)
LT3 Acroleine
LT4 Methylisocyanate
LT5 (not transported in bulk)
LT6 (not transported in bulk)
GE0 Not considered in a QRA
GE1 Methylmercaptan **)
GE2 n-Butane
GE3 n-Propane
GT0 Not considered in a QRA
GT2 Ethylchloride ***)
GT3 Ammonia
GT4 Sulphur dioxide
GT5 Chlorine
*) When used Ior studies on waterway transport do not consider sinking
and dissolving.
**) Just consider Ilammability.
***) Take substance properties at 293 K.
4.3 Road transport
4.3.1 Transported substances
In order to speciIy the LOCs Ior accidents in the transport oI hazardous materials, inIormation
on the substances transported and the packaging oI substances are essential. Eor the transport by
road in the Netherlands this inIormation is unIortunately not automatically collected and stored,
in contrast to the transport by rail or on inland waterways. However, in recent years inIormation
has become available by means oI visual registrations made at a signiIicant number oI locations.
Eor most oI the main roads in the Netherlands, inIormation on the annual transport and
composition oI transport (number oI transport movements per substance category) is given in
|AVIV97|. As the typical observation period lasted only 8 hours, the estimated transport streams
are to a certain degree uncertain and give merely an indication oI the magnitude oI the
transport. Essentials oI the adopted working procedures are given in |AVIV97|.
Basic Data 4.4
4.3.2 Determination of location specific outflow frequencies
In a QRA Ior road transportation the outIlow Irequency has to be speciIied. The outIlow
Irequency is deIined as the Irequency oI an accident with a hazardous substance transport unit
where at least 100 kg oI the transported substance is released. Generic values (per transport unit
km) Ior diIIerent road types are given in chapter 3.2.
In a detailed QRA the use oI location speciIic outIlow Irequencies Ior main roads is
recommended. As the number oI hazardous material accidents is too small Ior deriving a
location speciIic outIlow Irequency, it is current practice in the Netherlands to assume that the
outIlow Irequency is a linear Iunction oI the injury accident Irequency. The outIlow Irequency
Ior a particular road section may then be determined by multiplying with the ratio oI the
(location speciIic) injury accident Irequency to the average injury accident Irequency Ior that
road type in the Netherlands.
The procedure Ior the derivation oI a location speciIic outIlow Irequency involves the
determination oI:
The number oI accidents.
The amount oI traIIic.
The location speciIic injury accident Irequency.
The ratio oI the location speciIic injury accident Irequency to the average injury accident
Irequency.
The location speciIic injury accident Irequency should be derived as the quotient oI the number
oI injury accidents and the amount oI traIIic (number oI motor vehicle kilometres) in a
corresponding period using similar selections. Road data, number oI motor vehicles, and motor
vehicle accidents are registered in databases maintained by the TraIIic Research Centre AVV oI
the Ministry oI transport, Public works and Water Management:
Characteristics oI main roads can be retrieved Irom the database WEGGEG.
The number oI accidents may be retrieved Irom the database VOR. A subset oI this database,
IMPULS, contains all accidents on the main roads on a hectometre basis.
Amounts oI traIIic Ior approximately 1000 sections oI the main roads can be obtained Irom the
database INWEVA.
Local authorities (Regional Directorates oI Rijkswaterstaat (the Department oI Public Works)
and Provincial and Community Road management) should be consulted to obtain relevant data
Ior other road types.
Average injury accident Irequencies (number oI injury accidents Ior every million kilometres
travelled by motorised road vehicles per year) Ior the Netherlands are derived Ior several road
types: motorways, secondary and tertiary roads outside built-up areas and major arteries inside
built-up areas |SWOV97|.
Notes:
1. Review oI the distribution oI the location oI accidents along a certain road section may
indicate possible concentration points. Whether this is signiIicant can be statistically checked.
A method, which assumes that the accidents are binomially distributed, is outlined in
|AVIV94|.
Basic Data 4.5
2. In the derivation oI a speciIic outIlow Irequency a comparison has to be made between the
location speciIic accident Irequencies and an average injury accident Irequency. lt should be
ascertained as to whether the accidents and amount oI traIIic used in deriving the location
speciIic accident Irequencies reIer to the same selection as used in deriving the average
accident Irequency.
4.4 Railway transport
In order to speciIy the LOCs Ior accidents on private sidings, shunting yards and railways,
inIormation on the transport stream is required. This concerns speciIically the number oI
atmospheric and pressurised tank cars per period (usually one year) and the identiIication oI the
substances involved. The inIormation source diIIers by activity:
Eor a private siding, transported goods are usually only used by this establishment. ThereIore
the substances involved are roughly known. Eurthermore, as the private siding is a part oI
the establishment, the permit oI the establishment contains inIormation on the number oI cars
and the substances;
A shunting yard is an establishment with its own permit. This permit should contain all the
required inIormation on the number oI tank cars and substances;
The main railway operator in the Netherlands, NS Cargo, keeps records oI the transport
(number oI cars combined with substances) oI railway tank cars on the Dutch railway
network. This inIormation is not publicly available, but has to be ordered Irom NS Cargo.
In section 3.3 generic values Ior outIlow Irequencies are given Ior private sidings, shunting
yards and the railway network. It is strongly recommended to apply these generic values as there
are no publicly accessible databases or reports available on the numbers and types oI accidents
Ior a speciIic private siding, shunting yard or part oI the railway network.
However, all accidents on the railways are reported to the Dutch Railways (also Irom shunting
yards). II the accident is severe enough, a detailed investigation takes place. The Dutch railways
keep record oI all reported accidents. lt is not easy, however, to retrieve suIIicient inIormation,
that will justiIy the use oI speciIic outIlow Irequencies Ior the location under consideration.
4.5 Inland waterway transport
4.5.1.1 Introduction
In order to speciIy the LOCs Ior accidents in the transport oI hazardous materials, inIormation
on the substances transported and the packaging oI substances are essential. Eor the inland
waterways in the Netherlands this inIormation is stored in the IVS90 system. In IVS90 a
particular transport movement is tracked Irom sender to destination. At a number oI locks, data
on the passages oI ships and their contents are registered in a standard Iormat.
Basic Data 4.6
The annual transport and composition oI transport on a speciIic section oI a waterway may
thereIore be retrieved Irom IVS90. In the next paragraphs the essentials oI the working
procedures are given.
4.5.1.2 1ransported quantities
To Iind the annual quantity transported per substance along a speciIic section oI a waterway the
local manager oI the waterway should be contacted. IVS90 is partitioned into 5 regions: Grote
rivieren, Zuid West, Midden Nederland, Noord en Zuid Oost. Every region is divided into a
number oI blocks. Eor a speciIic year the IVS90 standard records oI the nearest IVS-block
should be requested. Normally the most recent three-year period is suIIiciently accurate. On
querying these records the Iollowing conditions should be applied:
1able 4.3 Conditions for retrieving shipped quantities from IJS90-records
Item Condition Remark
Mass > 30 tonnes Standard LOCs Ior tank ships must apply
Cargo type 2 Code Ior AD(N)R substances
Vessel type 2, 4, 6, 8, 10, 12, 14, 16, IVS90 ship-type codes Ior transport in bulk (tanks);
18, 30-39, 49, 52-54 containers and solids in bulk are not usually
considered in a QRA
In IVS90 the substances are registered by their UN-number. A vessel may, however, contain
more than one tank and several diIIerent substances. The standard LOCs assume leakage oI one
tank. The Irequency oI impact per tank is, as a Iirst and rough approximation, given as the
inverse oI the total number oI tanks in the vessel. The analysis should result in a list oI UN-
numbers (substances) and the number oI vessel and tank passages. As such this may not be an
adequate input Ior a QRA. The number oI substances may be reduced to a practical number by
categorising, as described in paragraph 4.2 above. Also the relation between substance
(category) and vessel type should be speciIied. This is outlined in chapter 3.4.
It is possible to select Irom IVS90 speciIic sender-destination combinations in order to analyse
their contribution to the total risk. Destinations are represented in the IVS90 by a numerical
code. It is also possible to select a speciIic type oI vessel Ior analysis or Iocus on some other
aspect depending on the character oI the QRA.
When Iuture developments are important some use can be made oI Iorecasting studies like
|NEA97| and |NEA98|. However, Iorecasts on the level oI speciIic hazardous substances are
rarely available. Extrapolating Irom the Iorecast Ior all commodities or Ior vessel passages is not
the proper way but oIten the best attainable.
Basic Data 4.7
4.5.2 Determination of location specific accident frequencies
4.5.2.1 Introduction
The location speciIic accident Irequency as an adequate input Ior a detailed QRA should be
derived as the quotient oI the number oI vessel accidents with serious damage and the amount
oI traIIic (number oI vessel kilometres) in a corresponding period. Accidents on waterways are
registered in the ONOVIS database maintained by the TraIIic Research Centre AVV oI the
Ministry oI Transport, Public works and Water Management. The amount oI traIIic
(passages/year) may be retrieved Irom IVS90. The accident Irequency may thereIore be deduced
Irom ONOVIS and IVS90 together. In the next paragraphs the essentials oI the working
procedures are given.
4.5.2.2 Accidents
Accidents on inland waterways in the Netherlands are reported by a number oI parties (river
police, regional authorities) and are nationally registered in the ONOVIS database. Eor a speciIic
section oI a waterway, data on the number oI vessels involved in accidents and the damage may
be retrieved Irom ONOVIS oI the TraIIic Research Centre AVV. Eor a QRA the most recent
5 years should be considered. A check should be perIormed as to whether major modiIications
oI the waterway were undertaken or other circumstances changed in that period. On querying
the records oI ONOVIS the Iollowing conditions should be applied:
1able 4.4 Conditions used ]or retrieving accident data on the main waterway from ONOJIS
records
Item Condition Remark
'Zijvaartaanduiding 'geen Selects accidents on main waterways and not
(branch indication) (none) the accidents in harbours or branches
'Shiptype 69 Selects accidents with working barges
(type oI vessel)
'scheepsschade omschrijving 'zwaar Selects accidents with serious damage. Serious
(description oI ship damage) (serious) damage is considered as an event where release
is a possible consequence
'scheepsschade omschrijving >'geen Selects all damages. When no serious damages
(description oI ship damage) (none) are encountered, the Irequency oI a serious
damage may be estimated as a Iraction oI all
damages (Table 4.6)
The items 'vaarwegnummer (waterway number) and 'kilometre indicate the location oI the
accident. A description oI these locations may be Iound in |AVV97a|. A crosscheck on the
location with the Iield 'accident_description should always be perIormed.
Eor gas tankers a more severe damage class (Iormerly called 'class 4) is needed to produce a
release oI product. Only when independent data on this damage class are lacking may deIault
values Ior the damage class Iraction be taken Irom Table 4.5 |AVIV93|.
Basic Data 4.8
1able 4.5 Damage class relevant to gas tankers as a fraction of total serious damages (only
to be used when lack of other data)
Navigability class Gas tanker damage class as a
(CEMT) fraction of total serious damages
4 0.14
5 0.19
6 0.32
When no accidents with serious damages are encountered (and only then, as this is a rough
approximation!), the number oI accidents with serious damage may be estimated as a Iraction oI
the total number oI accidents with damage (all-damage classes). This Iraction is derived Irom
|AVIV93| and is given in Table 4.6.
1able 4.6 Serious damage as a fraction of all damages (rough approximation in lack of other
data)
Navigability class (CEMT) Serious damages as a
fraction of all damages
4 0.16
5 0.27
6 0.34
Summarising:
1. Eind the number oI serious damages in the period considered Irom the ONOVIS database.
2. When no serious damages occurred in the period considered, Iind the number oI all-damage
classes Irom the ONOVIS database and multiply this by the Iraction Irom Table 4.6.
3. Eind the number oI the more severe damages (relevant to gas tankers) either Irom the
database or apply the Iraction Irom Table 4.5 to the number resulting Irom step 1 or step 2.
When the result oI Step 1 or 2 is divided by the number oI years and by the number oI
kilometres considered this gives the damage density (per km.yr) Ior single hull and double hull
vessels. When the result oI Step 3 is divided by the number oI years and by the number oI
kilometres considered this gives the damage density (per km.yr) Ior gas tankers. To calculate the
damage frequency (per vessel km.) the amount oI traIIic needs to be taken into account.
4.5.2.3 Amount of 1raffic
The amount oI traIIic (passages/year) may be retrieved Irom IVS90 or Irom the annual
publications oI the TraIIic Research Centre AVV |AVV97b|. Recreational traIIic is not
considered.
4.5.2.4 Determination of accident frequency
The location speciIic accident Irequency, as an adequate input Ior a detailed QRA, is derived as
the quotient oI the number oI accidents with serious damages and the amount oI traIIic (number
Basic Data 4.9
oI vessel kilometres) in the corresponding period. A Irequently encountered problem is whether
a speciIic section oI the waterway can be regarded as nautically homogeneous, that is whether it
may be assumed that the accidents that occurred at speciIic locations on the waterway section
might just as well have happened anywhere else along that section. A review oI the distribution
oI locations oI accidents along the waterway may indicate possible concentration points. Also, a
nautical expert may judge whether (and which) speciIic local Iactors are present that may
explain an increased accident rate. II so, the accidents (or some oI them) have to be considered
speciIic Ior that location. II not ('the accidents might as well have happened anywhere else
along the section) the accidents should be considered representative Ior the total length oI the
nautically homogeneous section. Erom studies, it appears that passages involving bridges or
locks bring about very speciIic accidents that are not representative Ior adjacent sections oI the
waterway, and these accidents are thereIore assigned to a speciIic location.
When risk-reducing measures are considered and evaluated in a detailed QRA, a model is
needed which links the characteristics oI the waterway and the amount oI traIIic to the accident
Irequency. The TraIIic and Accident model (V&O model |MSCN95|, |AEA95|) oI the TraIIic
Research Centre AVV perIorms that task. Based on traIIic simulations and observed correlations
in waterway- and accident-characteristics, this model calculates the eIIect on the accident
Irequency oI variations in, Ior example, navigable width or amount oI traIIic and composition.
Eor the purpose oI calculating the risks oI Iuture developments, the V&O-model may also be
applied in studying the eIIect oI Iorecasted changes in amount oI traIIic and traIIic composition.
The application oI the V&O model, however, requires some nautical expertise.
4.6 Pipeline transport
The substance transported through a selected pipeline is usually well known. Nearly all
pipelines are designed and built Ior the transport oI one substance only. Eor the transport oI this
product only a permit is usually given. There are, however, a Iew pipelines through which more
substances than one are transported (not at the same time). This permit oI this pipeline should
provide the required inIormation.
Paragraph 3.4.3 presents the generic values Ior the outIlow Irequencies oI pipelines. The use oI
these values is strongly recommended. This is because the number oI accidents Ior a speciIic
location is usually very low and it takes quite a number oI years to come up with a reliable
Irequency.
Basic Data 4.10
5. REFERENCES
|ADNR| Reglement voor the vervoer van gevaarlijke stoIIen over de Rijn (Regulations Ior
the carriage oI dangerous substances over the Rhine), SDU, The Hague, 1998
|AEA95| Handleiding risicobepalingsmethodiek en doorrekenen maatregelen (Manual Ior
risk determination method and the consequences oI measures), AEA Technology
Netherlands B.V., The Hague, 1995
|AVIV93| Aandachtspunten HooIdvaarwegen (Points oI interest Ior Major waterways).
AVIV, Enschede,1993.
|AVIV94| Eundamenteel onderzoek naar kanscijIers voor risicoberekeningen bij
wegtransport gevaarlijke stoIIen. (Eundamental research on probability numbers
Ior risk assessment oI road transport) Enschede: AVIV, 1994.
|AVIV94| Handleiding risicoberekening wegtransport gevaarlijke stoIIen. Bepaling
Iaalkansen. (Manual Ior risk assessment oI road transportation oI dangerous
substances. SpeciIication oI chances oI Iailure) AVIV, Enschede, 1994.
|AVIV95| Methodology Ior categorising substances in risk calculations oI the transport oI
hazardous materials. SaIety on waterway transport project S3b. AVIV, Enschede,
1995.
|AVIV99| Methodology Ior categorising substances in risk calculations oI the transport oI
hazardous materials. AVIV, Enschede 1999 (to be published).
|AVIV97| Risico`s wegtransport gevaarlijke stoIIen (Risks oI road transport oI dangerous
substances) AVIV, Enschede, 1997.
|AVV| Wegwijzer voor de binnenscheepvaart (Handbook Ior inland navigation). AVV,
Rotterdam
|AVV97a| Vaarwegen in Nederland (Waterways in the Netherlands). Ministry oI Transport,
Public Works and Water Management, Advisory service TraIIic and Transport,
Rotterdam, 1997.
|AVV97b| KerncijIers hooIdvaarwegen. (Key data main waterways) Ministry oI Transport,
Public Works and Water Management, Advisory service TraIIic and Transport,
Rotterdam 1997.
|Buis| Structuurschema Buisleidingen (Structure Scheme Ior Pipelines)
Kamerstuk 11; 173 53, nrs. 1-2
|CPR14E| Committee Ior th eprevention oI disasters,Methods Ior the calculation oI physical
eIIects (the 'Yellow Book), The Hague SDU 1997, PGS 2.
|IPORBM| IPO RisicoBerekeningsMethodiek IPO (IPO Risk Calculation Methodology).
AVIV, Enschede, 1997.
|MSCN95| Kok M., Tak C. van der, Gebruikershandleiding van het verkeers- en
ongevalsmodel versie 1.2, (Users Guide to TraIIic and Accident Model Version
1.2), MSCN Wageningen, 1995
|NEA97| Beneden rivieren also hooIdtransportas, Scheepvaartprognoses 2015. (Tidal
rivers as main transporters, Shipping Iorecasts 2015) NEA, Rijswijk 1997.
ReIerences 5.1
|NEA98| Scheepvaartprognoses Zeeburg en Lekkanaal (Shipping Iorecasts Zeeberg and
Lekkanaal). NEA, Rijswijk, 1998.
|NEN92| Eisen voor stalen transportleidingen (Requirements Ior steel transport lines).
NEN 3650,1998.
|PROT98| Protocol QRA seagoing vessels: a project proposal, Ministry oI Transport, Public
Works and Water Management, 1998.
|RIVM99| Guideline Ior Quantative Risk Assessment.
|SAVE88| Risico`s van het bulkvervoer van brandbare en giItige stoIIen over het water
(Risks oI bulk transport oI Ilammable and toxic substances by water), SAVE,
Apeldoorn, 1988.
|SAVE95| BasisIaalIrequenties voor het transport van gevaarlijke stoI over de vrije baan
(Basic Iailure Irequencies Ior transport oI dangerous substances on open rail
track). Report 951675-556, produced Ior N.V. Nederlandse Spoorwegen. SAVE,
Apeldoorn, 1995.
|SAVE95a| BasisIaalIrequenties voor het transport van gevaarlijke stoIIen per spoor
(emplacementen). (Basic Iailure Irequencies Ior transport oI dangerous
substances in rail yards). Report 951599, SAVE, Apeldoorn, 1995.
|SAVE95b| Risico`s vervoer gevaarlijke stoIIen, deelproject A74: buisleidingen; deel 1 (Risks
oI the transport oI dangerous substances, project A74: pipelines; part l). Report
951250- 586, produced Ior workgroup IPO-A74, SAVE, Apeldoorn, 1995.
|SAVE97| Handleiding/protocol voor het uitvoeren van een QRA voor
goederenemplacementen. (Handbook/protocol Ior the carrying out oI a QRA Ior
goods yards). Report 971071 - A27, produced Ior the Ministry oI Transport,
Public Works and Water Management. SAVE, Apeldoorn, 1997.
|SWOV97| Risico`s onderscheiden naar wegtype (Risk diIIerences between road types). R-
96-62. SWOV, Leidschendam, 1997.
|V&W89| Nota risico`s van het bulkvervoer van brandbare en giItige stoIIen langs de
vaarweg Rotterdam-Duitsland, Bijlage 1 (Note on the risks oI bulk transport oI
Ilammable and toxic substances along the waterway Rotterdam-Germany,
Annex 1) The Hague, Ministry oI Transport, Public Works and Water
Management, 1989
|VeVoWeg96| Eindrapport deelnota 3; handreiking risicobepalingsmethodiek externe veiligheid
vervoer gevaarlijke stoIIen over de weg & voorbeeldstudie. Project Veiligheid
Vervoer over de Weg. (End report part 3: Handbook Ior risk determination
method Ior external saIety oI transport oI dangerous substances by road and
example study. Project SaIety oI Transport by Road) The Hague: Ministry oI
Transport. Public Works and Water Management, 1996.
|VNG98| Handreiking externe veiligheid vervoer gevaarlijke stoIIen, (Handbook on
external saIety in the transport oI dangerous substances) Association oI Dutch
Communities; Ministry oI Transport, Public Works and Water Management;
Ministry oI Housing, Spatial Planning and Environment; Ministry oI Internal
AIIairs, The Hague 1998
ReIerences 5.2
|VROM84| Zonering langs hoge druk aardgastransportleidingen (Zoning along high pressure
natural gas transport lines) Circulaire van de minister van VROM (Circular oI the
minister oI Housing, Spatial Planning and Environment), 26 November 1984
ReIerence DGMH/B no. 0104004.
|VROM91| Circulaire voor de zonering langs K1-, K2-, en K3-vloeistoIleidingen. Bijlage bij
brieI van minister van VROM van 24 april 199l. (Circular Ior the zoning along
K-1, K-2, and K-3 liquid lines. Annex to the letter oI the minister oI Housing,
Spatial Planning and Environment oI 24 April 1991) reIerence
DGM/SR/1221254.
|VROM95| Risicobenadering voor NS-goederenemplacementen. (Risk approach Ior NS
Ireight yards). VROM-circulaire (Circular oI the Ministry oI Housing, Spatial
Planning and Environment), 1995.
|VVoW95| Systematiek voor indeling van stoIIen ten behoeve van risico-berekeningen bij
het vervoer van gevaarlijke stoIIen. Project Veilig Vervoer over Water;
Deelproject S3b. (System Ior the categorisation oI substances with respect to risk
calculations Ior the transport oI dangerous substances). The Hague: Ministry oI
Transport. Public Works and Water Management, 1995.
|VW98| Quantitative risk assessment Ior transportation oI hazardous materials.
IPO Risk Calculation Methodology. Road-Rail-Water-Pipeline.
The Hague, Ministry oI Transport. Public Works and Water Management, 1998.
|WL95| Kok M., Klopstra D., Jong J.H. de, Veiligheid Vervoer over Water: Modellering
van maatregelen (SaIe Transport by Water: Modelling oI measures),
Waterloopkundig Laboratorium (The Hydrodynamic Laboratory), DelIt 1995.
ReIerences 5.3

Purple Book 2

Uploaded by

Copyright:

Available Formats

Purple Book 2

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Purple Book 2

Uploaded by

Copyright:

Available Formats

What are the three subjects that require additional research according to the report?

What are the three subjects that require additional research according to the report?

What is the difference between risk calculations for transport and stationary installations mentioned in the report?

What is the difference between risk calculations for transport and stationary installations mentioned in the report?

Guideline for quantitative

all grid cells

High speed tracks (>40 km/hr)

Low speed tracks (40 km/hr).

Pipeline on dedicated route Ior pipelines

NEN 3650/RTL new: pipeline built in compliance with NEN 3650

Description oI the transport units (characteristic inventory)

Description oI the transport route (road type, obstacles present)

Description oI the ignition sources

Properties oI transported (representative) substances

Terrain classiIication oI the surroundings oI the transport route

Population present in the surroundings oI the transportation route

release oI the complete inventory.

instantaneous release oI the complete inventory oI the transport unit.

continuous release Irom a hole with an eIIective diameter oI 50 mm (2 inches).

description oI the transport units (characteristic inventory);

description oI the transport route;

description oI the shunting yard or the private siding;

description oI the activities on the shunting yard or the private siding;

description oI the ignition sources;

properties oI transported (example) substances;

terrain classiIication oI the surroundings oI the transport route;

leakage Irom a 3" hole in the tank car;

rupture oI the tank car.

the sets oI points are manually operated;

the maximum speed is 30 km/hr;

level crossings with roads;

not equipped with signs.

Description oI the transport units (characteristic inventory)

Description oI the ignition sources

Properties oI transported substances

Terrain classiIication oI the surroundings oI the waterway

Population present in the surroundings oI the waterway

type oI containment oI the substance e.g. type oI ship or tank

substance properties e.g. chemical, physical, toxicological

Single hull vessel

Double hull vessel (transport at room temperature or reIrigerated transport)

overall width oI the waterway

navigable width oI the waterway

the distribution oI the traIIic on it

N-vessels (Rn. 331.000)

C-vessels (Rn. 321.000)

G-vessels (Rn. 311.000)

Dry bulk single hull vessels (Rn 110.000)

Dry bulk double hull vessels (Rn. 110.288)

Sinking liquids are not considered in the QRA.

Description oI the transport system (diameter, location emergency valves);

Description oI the transport stream (substance, Ilow rate);

Description oI the ignition sources;

Properties oI transported substances;

Terrain classiIication oI the surroundings oI the transport route;

Population distribution in the surroundings oI the transportation route.

Description oI the transport route (location, type oI route, obstacles present)

Description oI the transport units (type oI unit, characteristic inventory)

Description oI the ignition sources

Properties oI transported (representative) substances