Professional Documents
Culture Documents
Editor
A.Z. Keller, Department of Industrial Technology and Management,
University ofBradford, U.K.
Aims and Scope. Fundamental questions which are being asked these days of all
products, processes and services with ever increasing frequency are:
What is the risk?
How safe is it?
How reliable is it?
How good is the quality?
How much does it cost?
This is particularly true as the government, industry, public, customers and society
become increasingly informed and articulate.
In practice none of the three topics can be considered in isolation as they all interact
and interrelate in very complex and subtle ways and require a range of disciplines
for their description and application;, they encompass the social, engineering and
physical sciences and quantitative disciplines including mathematics, probability
theory and statistics.
The major objective of the series is to provide a series of authoritative texts suitable
for academic taught courses, reference purposes, post graduate and other research
and practitioners generally working or strongly associated with areas such as:
Safety Assessment and Management
Emergency Planning
Risk Management
Reliability Analysis and Assessment
Quality Assurance and Management
Special emphasis is placed on texts with regard to readability, relevance, clarity,
applicability, rigour and generally sound quantitative content.
The titles published in this series are listed at the end of this volume.
Quantitative Risk Assessment
of Hazardous Materials
Transport Systems
Rail, Road, Pipelines and Ship
by
MICHEL NICOLET-MONNIER
PSI. Paul Scherrer Institute. Villigen. Switzerland
and
ADRIAN V. GHEORGHE
ETHZ. Swiss Federal Institute ofTechlWlogy. Zurich. Switzerland
PREFACE xm
FOREWORD xvn
ACKNOWLEDGEMENTS XIX
LIST OF ABBREVIATIONS XX
In view of the above mentioned, a joint project, PPR&S (Polyproject on "Risk and
Safety of Technical Systems) was launched by ETHZ (Swiss Federal Institute of
Technology, Ziirich), together with the participation of the following institutions:
• Paul Scherrer Institute, PSI (ViUigen, Switzerland)
• EAWAG (Diibendorf, Switzerland)
There arc a number of national and international efforts to deal with risk assessment
at the regional level. The "Poiyproject on Risk and Safety of Technical Systems". at
ETHZ, took the initiative to research on various aspects related to regional risk
assessment and safety management A series of basic questions were asked, and answers
had to be given; they had to comprehend the following main issues:
what is integrated area risk assessment and safety management
how to define a region/area for study
type of activities and targets at risk
the need for risk impact indicators
the need for a comprehensive methodology.
Regional risk assessment and safety management seems to be a medium which helps
to integrate people. issues and decisions in area risk assessment
PPR&S is the discursive description of knowledge in addition to the development of
interdisciplinary and practical methods for the application of risk management for a
large variety of technological systems (e.g .• rail and road transportation of dangerous
goods, chemical plants, nuclear power plants, biotechnology, landfill, etc.). The scope
of these applications is to be presented as a book series. It is intended to be an integrated
PREFACE xv
methodology has been highlighted in different occasions during the time of the
Polyproject
Some of the lessons we have learned in the course of the PPR&S, when dealing with
above issues, are:
• When getting involved in a regional risk assessment do not take a simplistic
approach.
• On regional risk assessment, avoid excluding political or human interactions at
all stages. Develop a risk triplex, namely: "safety culture, environmental aware-
ness, and emergency culture".
• When running a research or a case study on regional risk assessment do not en-
tirely rely on a self-organizing effect within the project, or case study. A strong
interactive project management framework is needed from the beginning.
• Due to technical accidents or natural disasters, or their synergism, there is an
emerging need for national and international organized research and activities in
the above field Risk from normal operation or from accidents may have some
transboundary effects.
• Legal issues at the local level or national level have to be solved and harmonized
before any implementation of a comprehensive regional risk assessment meth-
odology can take place.
• There is a need for specialized databases; their systematic use will reduce the
uncertainty of the results.
• Recent advancements in information and telecommunication technologies (GIS,
ISDN), multimedia, virtual reality, neural networks, could play an important
contribution to the modeling of various risks.
• In regional risk assessment, all risks (local, regional, global) should be taken into
consideration.
• Safety culture, public participation and risk communication are relevant issues in
the overall landscape of the regional risk assessment process. Emergency culture,
preparedness, and planing is an integral part of regional safety management
As compared with similar projects in the world (e.g., the UN Inter-Agency on Risk
Assessment and Safety Management for Large Industrial Complexes), the present work
brought new answers to this interdisciplinary subject. WoIk done within the PPR&S is
complementary to the numerous activities developed recently in Switzerland.
Further information on the Polyproject and its publication series can be obtained
from:
Polyproject "Risk and Safety of Technical Systems"
ETH-Center Phone: +(41) 1 6322356
CH-8092 ZUrich Fax: +(41) 1 632 1094
Switzerland
FOREWORD
In recent years, the community has become increasingly aware of the risks of locating
hazardous industries near heavily populated environmentally sensitive areas. This new
awareness could lead to a novel approach to safety planing for hazardous industries,
looking at the problem from the point of view of integrated regional risk assessment,
which should include beside the risks arising from natural events (like earthquake,
flood, forest fires, etc.) also the risks arising from processing plants, storage and trans-
portation of dangerous goods.
The purpose of this volume is to highlight the main procedures for assessing the re-
gional risks resulting from dangerous goods storage, and transportation by means of
different technical systems (i.e., road, rail, ship, and pipeline).
The information contained in this book is based on a wide range of references and
studies. The main procedural steps involved in quantitative risk analysis for transporta-
tion systems are supported by relevant methods of risk assessment recognized on an
international level. The present book gives an overview of the criteria and guidelines
applicable to the implementation of risk assessment and management at different stages.
Chapter 1 describes the environmental and safety factors to consider, when perform-
ing a transportation risk analysis for a region. Chapters 2 presents risk definitions and
the methodology for analyzing transportation risks in a complex area. Chapter 3 pres-
ents general information about truck accidents and thei!" consequences, and reviews the
risk presented by road tunnels. The BUW AL methodology for assessing the risks of
hazardous materials transportation by road is explained in detail. Finally new develop-
ments in traffic and vehicle control are discussed. Chapter 4 deals with transportation
of hazardous materials by rail. The U.S and British experience is reviewed. information
concerning the Swiss Federal Rail Network is presented, then the BUWAL methodol-
ogy for assessing the risk of rail transportation is described and the findings of a risk as-
sessment study by the SBB are finally summarized. Chapter 5 is more concerned with
the assessment of transportation risks on water ways. Chapter 6 furnishes a descrip-
tion of the transport pipelines for natural gas and petroleum products and describes the
situation of Switzerland. Some information on storage and handling as well as informa-
tion concerning codes of practice and international regulations can be found at this
place. Chapter 7 presents a compilation of statistical data related to accidents and dan-
gerous goods' movements. This information is related more specifically to the case of
Switzerland Chapter 8 is devoted to the description of data bases and computer sup-
port for risk assessment. The content of several data bases and programs is presented
and the source of availability mentioned. Special literature and topics concerning GIS
applications, and other risk assessment methods are reported at this place. Chapter 9
deals with integrated approaches for regional risk assessment and safety management
with special emphasis to the transportation of hazardous materials. Chapter 10 presents
several relevant case studies and miscellaneous information.
XVIII FOREWORD
In Chapters 3 and 4, dealing with rail and road transportation, we have summarized
the method for assessing transportation risks, which is recommended by the Swiss Fed-
eral Office for Environment, Forestry and Landscape, BUWAL.
Due to the time and manpower imparted to the preparation of this book within the
framework of PPR&S, it was not possible to assess in detail the risks presented by the
rail and road transportation of hazardous materials in Switzerland Much to our regret it
must be stressed that the Swiss statistics concerning road tankers' and trucks' accidents
are not suitable for probabilistic risk assessment Due to this situation we have re-
nounced to assess globally the transportation risk for Switzerland on the basis of a prob-
abilistic approach. However the methodological approach to consider has been de-
scribed with much details. The task of the authors was to review the state of the art
concerning the analysis and assessment of transportation risks for dangerous goods,
with a special issue concerning Switzerland. In that respect we hope that our contribu-
tion will be of some value to the reader interested by such topics.
Environmental impacts from accidental releases ofpoUutants into air, water, and soil
were mainly treated in Volume II of Integrated Regional Risk Assessment, by the same
authors, published recently by Kluwer Academic Publishers, Dordrecht, The Nether-
lands.
It should be noted that the risks of nuclear fuel and radioactive wastes transportation
have not been treated here at length. This is because nuclear risk assessments are cur-
rently carried out at a higher level than that used for other facilities (e.g., in Switzerland
by the HSK, Villigen) and the information would be available for use in integrated risk
assessment at community level.
This book shall be valuable to students, engineers, and scientists in charge of dew'l-
oping new methodologies for !.Iazard analysis and risk assessment for transportation
systems; practitioners active in the field of environmental protection; local or govem-
mental Authorities in charge of implementing environmental risk impact assessment
procedures and guidelines.
ACKNOWLEDGEMENTS
The authors wish to express their sincere gratitude to Professor Wolfgang Kroger,
Chairman of the Executive Committee of the "Polyproject, Risk and Safety of Technical
Systems (PPR&S)", ETHZ - Swiss Federal Institute of Technology, Ziirich, for his
guidance and critical reviews during the different phases of this work.
We are also greatly indebted to all people, who through their support and many
valuable suggestions for corrections and improvements of the manuscripts, helped us to
finalize this work. We wish also to acknowledge more specifically the following indi-
viduals:
Prof. R. Hatter, Vice-President Research, ETH ZUrich
S. Chakraborty, HSK, Villigen, Switzerland
Dr. Hans-Jorg Seiler, Project Manager for the PPR&S
Prof. J. Schneider, Institut fUr BaustatiklKonstruktion, ETH, ZUrich
Mr. H.A. Men, EBP, Basler Ernst & Partner Ingenieuruntemehmen, ZUrich
Prof. Dr. B. Bohlen,former DirectorofBUWAL, Bern
H.R. Wasmer, Deputy Director, EA WAG, Dabendorf
Prof. K. Hungerbuehler, Institut fUr Technische Chemie, ETHZ
Dr. H.P. Alder, Staff Nuclear Energy and Safety Research Department, PSI, Villigen
Dr. R. Kieselbach, EMPA, Dabendorf
Dr. H. Kiinzi, Konzernsicherheit und Umweltschutz, Hoffmann-La Roche AG, Basel
Mr. K. Cassidy, Head Major Hazard Assessment Unit, Health and Safety Executive,
London.
One of us (A.G.) would like to express special consideration and high appreciation
to Mrs. Fran~ise Bordier for her exquisite support and distinguished encouragement in
his professional activity while in Switzerland. Finally, in the preparation of this book,
the authors are greatly indebted to Mrs. I. Kusar (PSI), who skillfully prepared the
drawings and pictures for illustrating the manuscript.
LIST OF ABBREVIATIONS
It was not the purpose of the present book to analyze in depth the risk of transportation
of dangerous goods in Switzerland, but instead, in the framework of the Poiyproject
Risk and Safety of Technical Systems (PPR&S), we concentrated mainly on applied
risk analysis techniques and special attention was also given to highlighting the
methods and to complement them with examples and cases studies, to present general
infonnation and, as far as available, international as well as Swiss national statistical
data on accident frequencies, accident speed, spill frequencies, and other matter related
to the risk analysis process of hazardous materials transportation systems, involving
road, rail, ship and pipelines, to be viewed in the context of integrated regional risk
analysis.
In the first two chapters we present important definitions concerning risk, and
explain in a general way the methodology to apply for assessing the risk of hazardous
materials transportation. The Chapters 3 through 6 are treating more in detail specific
topics of road, rail, ship and pipeline transport systems. The reader will find various
st;ltistics and infonnation on accident, releases of hazardous materials, and conse-
quences of fire and explosion. Chapter 7 is more concerned with infonnation and
various statistics on dangerous goods movements in Switzerland. Chapter 8 describes
data bases of interest, and in Chapter 9 we also present some concepts of integrated risk
assessment and safety management at regional level related to hazardous materials
transportation. Finally in Chapter 10 we present some case studies showing the
methodology followed by different companies or authorities in Switzerland and in other
countries, for assessing transportation risks of hazardous materials. These case studies
present further methodological details and thus complement the topics covered in the
other chapters.
In this book we also describe the methodology recommended by the Swiss Federal
Office for Environment, Forestry and Landscape (BUWAL)I for assessing the risk of
dangerous goods transportation by rail and road in Switzerland. The principles of the
Swiss methodology can easily be applied to countries other than Switzerland, under the
condition that the coefficients used in the equations be replaced with coefficients
reflecting the conditions prevailing in the country under consideration, because the
accident statistic and traffic conditions might differ considerably.
The transport and distribution of hazardous materials, such as petroleum products,
liquefied petroleum gases, chlorine gas, pesticides, chemicals/petrochemicals and radio-
active materials, inevitably involve the potential for incidents and accidents which may
2 CHAPTER 1
It should be noted that although this chapter deals mainly with the risks of
hazardous materials being transported, when considering the overall system, the risks of
choosing a particular mode of transport may well be dominated by the risks involved at
the loading/unloading facility. These risks can be calculated either as part of the fixed
facility assessment or as part of the transport assessment - whichever is the more
appropriate for the specific situation. It should also be noted that, for most forms of
transport of hazardous materials, there exists a comprehensive and extensive list of
international transport regUlations.
The results of a transport risk assessment are very specific to that particular assessment
and care must be taken to avoid drawing conclusions from a few studies and then
generalizing from these, i.e., rail may not always be safer than road, pipelines muy nat
always be safer than raj{ etc.
This section describes the procedures for analyzing and comparing alternative routes
for the transportation of hazardous materials on the basis of land use and environmental
safety. It is not intended to provide in-depth documentation of the assumptions and
processes implicit in the methodology. Rather, the purpose is to highlight the most
relevant procedural information and a concise description of the criteria that may be
applied for hazardous materials routing.
Factors that influence routing decisions, from an environmental safety viewpoint,
may be grouped into three inter-related categories (see box):
• Mandatory factors such as legal and physical constraints (e.g., topography of the area).
• Environmental and land use risk (including various hazards and their associated risks).
• Subjective factors that reflect community priorities and values which may not be easily
quantified (e.g., population distribution, special land uses, routing, and emergency
response).
4 CHAPTER 1
quantities of dangerous goods these companies transport and the route they use
when moving these commodities.
Road Accident Rate. The probability of a dangerous goods accident can be estimated
in two steps. First, the likelihood of an accident for any accident must be extracted from
historical data records for the road network, using national statistics or when missing
using the reported statistics of countries, where similar traffic conditions are prevailing.
The accidents must be matched with the volume of vehicle traffic, expressed as the
Average Annual Daily Traffic (AADT). For highways and other national roadways, a
representative accident rate can be derived from their physical characteristics, such as
number of intersections and average traffic volume. Accident rates can be computed for
a given road length (called a link) according to the formula:
. Accidents
ACCIdent rate =-----
Vehicle- km
(1.2)
It has been recognized, that among the factors that contribute to dangerous goods vehicle
accident, humon error ranks highest. The most common causes of vehicle accident are due
to excessive speed, following to close the preceding vehicle, non observance of rest-time
leading to driver overfatigue, and failure to observe traffic warnings.
6 CHAPI'ER 1
Rail Accident Rates. The selected method relies on available dangerous goods accident
and traffic infonnation to calculate an accident rate for each of the given rail links.
In Switzerland, statistics on rail accidents are prepared on an internal basis by the
Swiss Federal Railroad (SBB/CFF). Historic accidents for cars carrying dangerous
goods can be identified for a given period and divided by the total number of dangerous
goods (DG) car movements over the same period to yield:
DG accidents per year Accidents
----------~~~---=----------
DG car movements per year Car movements
li!!K: 00. here. means dangerous goods.
The inference is that the" greater the accident rate for a given link, the greater the
likelihood of a dangerous goods rail car being involved in a crossing accident, collision
or derailment A typical value for the rail accident rate is 6.0 x ](}6 accidentlcar-km.
This approach assumes that the historic accident rate, recorded by link, predicts the
future accidents.
Emergency Response
Regional and national contingency plans for emergency preparedness and response
must be taken into consideration, when preparing the risk assessment study.
Risk Assessment
For assessing road and rail risks, the study should include the following areas:
a) Goods in Movement. For transport en route, the population at risk is twofold: (i)
people living near the road or railway, and (ii) the other users - passengers on
TRANSPORTATION RISK - ENVIRONMENTAL AND SAFETY FACTORS 7
passenger trains, or users of other road vehicles which might become involved in
a major accident
b) Temporary Stopover Points. These include railway marshaling yards, hucke-
pack loadinglunloading yards, lorry parking areas at motorway service points,
custom yards, etc.
c) Goods Handling. The loading/unloading process itself for trucks and road-
tankers.
Substances to Investigate
The study should in a first phase be limited to those substances which are transported in
large quantities and which have the potential to cause major accident risks from fue,
explosion and toxic release. Such hazardous substances belong to the following cate-
gories:
a) explosives
b) flammable gases
c) toxic gases
d) highly flammable liquids (with flash point < 21°C)
e) flammable liquids (with flash point between 21 °C and 55°C)
1) flammable solids (including substances liable to spontaneous combustion)
g) substances which become dangerous when wet
h) oxidizing substances including liquid oxygen
i) toxic liquids.
Thus among the substances listed above, priority should, in a first step, be given to
three groups of substances:
- those which appear to be potentially the most hazardous (for instance,
specifically liquefied toxic and flammable gases, with liquefied petroleum gas
(LPG), chlorine and ammonia)
- those moved in the greatest quantities by road or rail, according to goods
movement statistics (e.g., flammable liquids lique motor spirit, domestic fuel oil)
with a potential to take fIre or to contaminate surface and groundwaters.
- various explosives and ammunitions
Other substances can and do present risks (e.g., oleum, or fuming sulfuric acid).
However molten solids. acids. liquefied inert gases and caustic substances will usually
either be transported in small quantities, or have a limited range of action for putting
only few people at risk at anyone time.
Loading and unloading can be a source of potential risk. Loading is usually done at
sites which come within suitable regulatory regimes. It is therefore the unloading
(mainly at sub-notifiable sites) by road tankers of motor spirit. LPG and ammonia,
which should be studied in detail.
8 CHAPTERl
Harbors, unloading facilities, stopover points and marshaling yards present risks similar
in nature to those of fixed installations: - a (temporary) concentration of dangerous
substances at the vicinity of a local population surrounding it. In this case both
individual and societal risks should be estimated.
For road and rail en-route risks, it is the societal risk that becomes dominant because
the population at risk is constantly changing during the course of the day and the
individual may be only exposed to the hazard for a short time. However the total
number of individuals at anyone place potentially at risk from a release of a quantity of
hazardous substance may be larger than at a factory where such substance is produced
or used.
In the case of rail and road transport, the calculation of the societal risk is complex
because a route may expose a population in many towns and elsewhere, hence the risk
is spread over a wide area. There is also the group in transit of other road users, or
railway passengers, who might become involved in a major incident. A new concept of
"route societal risk" must be applied to consider the aggregated societal risk presented
by the transport of substances from one place to another.
An estimate of national risk from the carriage of a hazardous substance by rail or
road can be calculated from the route societal risk. scaling up by the volume of trade in
the given substance and length of route.
The accuracy of this scaling depends on the extent to which the routes chosen are
characteristic enough and truly representative of the conditions prevailing for the rest of
the country.
The risks from road transport of chlorine and ammonia are much lower than for
gasoline or LPG, but because ot the long-range effects if an accident of this kind shOUld
happen, there would be a very .. mall risk of an incident causing a very large number <1f
fatalities.
What concerns lorry stopover points, usually short-term parking is taking place in a
"safe place" with the vehicle usually in sight and within easy reach of the driver or a
competent person. The risk from chlorine and ammonia at a distance of 50 meters is
estimated to belO·6 (one in a million) per year of receiving a dangerous dose. This risk
at 50 meters would be consistent with the criteria for housing developments in the
vicinity of notifiable installations, though schools and nursing homes would need to be
built further away. For LPG the risk is somewhat higher. A distance of at least 120
meters would be recommended for certain kinds of housing developments and 170
meters for more vulnerable populations.
What concerns the loading/unloading of vehicles, there are a large number of
deliveries: for motor gasoline delivery is to numerous filling stations - usually urban,
but also on motorways and highways; For LPG, there is an expanding market and the
deliveries are made to filling stations respectively to a variety of sites, ranging from the
domestic to the industrial. Spills on delivery of motor gasoline is not infrequent, but
cases of ignition are seldom. Vent fire (i.e., ignition of the small quantity of vapor
necessarily escaping from vent) are more likely. The situation is different for LPG.
There the risks arise from errors in the filling process, or possible pull-away leaks
TRANSPORTATION RISK - ENVIRONMENTAL AND SAFETY FACTORS 9
(driver pulling away without disconnecting the hose). Ignition thereafter can and does
occur. It can result in a torch flame followed by a BLEVE. International experience
confinns the possibility of serious consequences. Less than half of the risk comes from
the unloading part of the transport process, the remainder comes from the installation
itself.
For ammonia the risk are somewhat larger, though the transport risk are only a small
proportion of the total risk from the site.
Hazard Classification. There are different classification scheme which may relate to
the substance itself or to qualitative differences (for toxic material) or to degrees of
hazard (for explosives). An alternative approach is to base the classification on the
hazard itself. Another important classification relates to the compatibility of chemicals
for adjacent bulk loadings. The hazard associated with compatibility are principally
evolution of gas and/or heat and evolution of hazardous materials. Such substances can
be flammable or corrosive.
The RID list (TABLE 1.1) indicates some of the chemicals which are regarded as
particularly hazardous in West Germany.
For certain chemicals which are transported by road special measures are required.
These include:
• Substances liable to polymerization (ethylene, butadiene)
• Substances carried fused
• Hydrogen peroxide
10 CHAPTER I
• Organic peroxides
• Sulfur trioxide
• Bromine
• Lead a1kyls.
In order to identify hazardous cargoes make use the UN list of hazardous materials
as basic check list. Special attention should be given to chlorine, ammonia, LPG
(liquefied propanelbutane gas), and other liquefied flammable gases; toxic gases;
flammable liquids and gases.
Figure 1.1. Method for seJecting dangerous goods in the study area
Identify the main modes of routes of transportation, and if possible identify also the
main origins and destination. Road, rail barge, ship, pipelines, as well as main transfer
facilities should be considered in the region of interest. Usually the origin and
destination of goods in transit from a foreign country to another one are known. The
check points are the main custom offices at the frontier. In Switzerland these are the
custom .offices of Basle, Chiasso, Geneva, etc. There the situation is even more
complicated for goods transported by ship on the Rhine river, because these are
(a)The Substance Hazard Index (SID) distinguishes chemical substances by their vapor pressures and toxicity
ratings (API, 1990). Use toxicity, flammability and reactivity to determine whether a chemical substance is
hazardous.
TRANSPORTATION RISK - ENVIRONMENTAL AND SAFETY FACTORS 11
complicated for goods transported by ship on the Rhine river, because these are
redistributed at their arrival in Basle between rail and road-tanker/trucks. No official
statistics exists for dangerous goods transports within Switzerland, beside the statistics
based on schedule of dues and not on dangerous goods classes that are established by
custom officers.
Physical Mandatory Factors: These may preclude a routing alternative and include
weight limitations on bridges, height restrictions on overpasses, inadequate shoulders
for breakdowns, extensive construction activities or inadequate parking and turning
spaces.
Laws and Regulations. They may apply to any routing alternative in prohibiting the
transport of hazardous materials through certain routes or structures (e.g., tunnels,
bridges). Local, state and national transport authorities should be consulted in all cases.
Such prohibited routes are obvious frrst cut alternatives to be eliminated.
In the USA, the DOT Research and Special Programs Administration RSPA guide
for routing analysis (RSPA8, 1988) allows routes to be excluded for physical or legal
reasons. Factors that are not easily quantified, such as the presence of schools or
hospitals that are not easily evacuated, or the presence of a reservoir, can be considered
subjectively. A recent notice of proposed rulernaking (FHW A, 1992) lists 13 elements
that should be considered in establishing highway routine requirements:
Population density
Type of highway
Type and quantities of hazardous materials
Emergency response capabilities
Result of consultation with affected persons
Exposure and other risk factors (such as distance to hospitals)
Terrain considerations
Continuity of routes
Alternate route
Effects on commerce
Delays in transportation
12 CHAPTER 1
Climatic conditions
Congestion.
Five criteria for routing analysis have been evaluated by Abkowitz et al. (1992).
these include: (a) minimizing shipping distance, (b) minimizing travel time, (c)
minimizing release-causing accident likelihood, (d) minimizing population exposure,
and (e) minimizing the product of accident likelihood and population. Of these, the first
two criteria minimize economic costs, and the latter three maximize safety. It was found
that the routes that minimize risk may be so sinuous that they can be economically
unfeasible, or at least impractical. Abkowitz et al. recommend that a routing analysis
consider combinations of factors and use different weighing factors to evaluate trade-
offs, such as travel time vs. minimum risk.
The risk, R j , for accident scenario i is a function of the scenario frequency and, Fj,
and the scenario consequence, Cj:
(1.3)
The product of the two terms is a relative risk indicator. Generally speaking, simple
risk indicators used for routing analysis can be considered as reduced forms of the more
complete quantitative risk equation, that will include other terms like probability of
leak, probability of given meteorological conditions, etc .. They do not provide as much
information regarding the safety of dangerous goods transport as do more complete
quantitative risk analyses. However, simple risk indicators can be useful in some
decisions involving two or more alternatives.
The overall environmental and land use safety criteria for route selection is that the
route which has the lowest risk value to surrounding people, property and the natural
environment should be selected. In this context, risk is determined in terms of the
cumulative combination of the probability of accidents and the consequences of such
accidents. These two elements of risk are dependent on the extent of population
exposed and number of properties or extent of natural environment ecosystems and the
accident rates. In general, routes with the smallest adjacent population as well as
accident rates, will have the lowest risk values.
It may also be insufficient to determine just the average risk over the whole route as it may
be necessary to determine the local risks adjacent to the more hazardous locations or the
more sensitive surroundings.
14 CHAYI'ER 1
TABLE 1.2. Potential impact area for different classes of hazardous materials
~: Determine the accident rates on a particular route. Ideally, the most reliable data
concerning accident rates would be those associated specifically with hazardous materials
transpartation tankers in terms of number of hazardaus mIlterials accidents per tankers. If
such infortnlltion is available then it should be used directly into probability estimlltions.
In 1tIIl1Iy cases, however, such infortnlltion is not readily available. It is usually necessary
therefore to rely on accident rates statistics for all vehicles and then to adjust these to
reflect the smlliler share of hazardous mIlterials in the traffic stream. The first step is
usually to obtain statistics from historical records of the total rates of accidents from all
vehicles, usually in terms ofAccidentslvehicles-km
~: Calculate the probability of an accident for any vehicle based on vehicle exposure.
The probability of any vehicle being involved in an accident of a specific segment is
calculated by multiplying the segment accident rate (from step I above), with the road
segment length (or amount of exposur,:). This probability is in terms ofAccidentslvehicle.
Step 3: Factor the probability statement for any vehicle to reflect the incidence of
hazardous mIlterials vehicles in the traffic stream. This is done by multiplying the proba-
bility figure from step 2 by the hazardous mIlterials accident factor (being the ratio of
hazardous mIlterials transport accidents I all vehicles transport accidents). This probability
is in terms of Hazardous Materials Accidents I vehicles.
Figure 1.2. Suggested sequence of steps for estimating the probability of transportation accidents.
where
PHMA : Probability of hazardous material accident
Note: Vehicle-Ian above refers 10 the total number of kilometers traveled by all vehicles for which
accident statistics are available. The length of link is identical with length of road segment
In cases where the values for hazardous material accidents could be obtained from
available statistics. then the following steps could be directly applied:
• Obtain the accidents statistics applicable to hazardous material tankers and convert to
Hazardous Materials Accidents/vehicle-km (Le .• per total number of km traveled by all
hazardous material tankers to which statistics apply);
• Obtain probability of hazardous material accident:
where
PHMA : Probability of hazardous material accident
Note: It is necessary in some justifiable circumstances 10 further introduce a correction factor that
reflect physical characteristics of the particular route segment which may increase the probability of
an accident on that particular route. The above equation should incorporate an allowance for this
factvf (F).
Risk Computations
The potential consequences (population and/or property) and accident probabilities for
each route segment are multiplied together to calculate the segment risk. The cumula-
tive summation across all route segments produces the total risk for the route. The
accident probabilities derived by the method'presented above should further be conver-
ted to a likelihood or probability of impact.
The probability of a release and a hazardous event occurring is computed using tools
such as event and fault tree analysis to incorporate factors such as:
• whether the load will be dislodged as the result of an accident
• extent of such load loss ensuing spillage
• effectiveness of any containment/emergency procedures
• likelihood of the spill or release reaching environmentally sensitive areas or
having an effect on people. buildings. etc.
order to achieve the coupling of events with the appropriate consequences in the overall
risk analysis
There are two basic approaches of estimating event probabilities: the first is direct
use of statistical data on failure of "whole systems or sub systems". This is sometimes
called the "historical approach". The second is to break down the event in its
contributory factors and causes. This process is pursued up to the point where data is
available, or can be obtained, for the contributory sub-events. The overall probability is
synthesized from the data for these various sub-events. The analytical approach will
inevitably be adopted where the historical data is inadequate or simply not available. It
may also be adopted where there is a cause for expecting a particular system or
component to have a sufficiently different failure probability from that indicated by
historical data.
Sometimes the analytical approach will not succeed in breaking the event down to
the point where data is available for all sub-events. In this case judgmental inputs may
be required and this has produced suspicion that the technique is not scientifically
rigorous. However, the philosophy here is consistent with the basic approach of
breaking down the judgment of how likely the event is into smaller areas of more
specific judgment which are within the realms of practical experience. The overall
outcome is then less sensitive to anyone particular judgmental input. An important
consideration in the analytical approach is how far to continue the breakdown of the
main event.
An advantage associated with the use of historical event data is that, where the
accumulated experience is relevant and statistically meaningful, the assessment will not
omit any of the significant routes leading to the event. The data already encompasses all
relevant contributory aspects including the reliability of equipment, human factors,
operational methods, quality of construction, inspection, maintenance, operation,
environment, etc. However, routes which may not be relevant to a specific case under
study will be included, resulting in an over-estimate - usually referred to as a
"conservative" estimate - of the chance of the event. The sample may also include older
equipment designed to lower standards.
Often the historical data is of such quality that, while undoubtedly relevant, it is not
considered adequate. In this case synthesis must be used, and the predicted event
probability should be tested against whatever experience exists, to judge whether the
different approached produce compatible predictions.
The chance of occurrence of an event is generally referred to by the term
"probability". This quantity may indeed be a probability and will therefore be dimen-
sionless. However, in many circumstances it will be expressed as a frequency of occur-
rence over a specified time interval, for example a year or a system lifetime. It may be
possible to relate such a frequency to finite number of occurrences over a sufficiently
long time, for example a large number of events over any relevant time scale they are
taken to be indicative of a degree of belief that an event will occur during a more
limited period. This distinction between "frequency" and "Bayesian" interpretations of
probability is generally considered academic in the context of risk analysis. However,
the distinction between dimensionless probability values and frequency values is
important where these numbers are to be used in combination. Because it is not always
TRANSPORTATION RISK - ENVIRONMENTAL AND SAFETY FACI'ORS 11
possible to be deftnite about which way a particular value will be expressed, some
authors use the tenn "likelihood" to express the chance of something happening, leaving
it open as 10 whether a "probability" or a "frequency" sho uld be used depending on the
circumstances.
Event Tru
Event trees follow initial causes through to possible outcomes. 1bey can be used to deal
with independent and coincident events, but ale particularly powerful for portraying
event sequences. Event trees are usually applied to binary state systems but can be used
for multi-<lutcome states. Although event trees are often used in their own right and
provide easy CJltension 10 q uantificatio n they are also useful in identifying sub-systems
which require faul t troe analysis. II is difficull 10 represent interactions between event
states and a separate troe will be required for each initiating event - the relationship
between different trees must then be considered very carefully. Outcomes ale related to
the specific cause being analyzed and could arise from other causes, which would not
be shown.
18 CHAPTER 1
In recent years a number of computer codes have been developed for event tree
analysis. TABLE 1.3 lists some of the more readily available computer codes.
The Figures 1.3 through Figure 1.5 present below different exemples of event trees:
a) for chlorine truck accident
b) for petrol road tanker accident, and
c) for a LPG tanker accident.
0·r 3
r
ill DiIIodged O'j
mm mm
No
(0.8) (0.2)
I I
Probability
0.4 0.1
Fatality
Radius(m) ~ b~~~~
pr o 1.3 1.8 2.8 4.0 5.3
I I I I
Probability 0.95 0.045 0.0035 0.0015
Small Release
Conto Small Initial Large Total Tank
<100 kg
Leak Leak Failure
Or No Damaae
Pipe Failure
/Hole
h 0000':"
(0.95) (0.05)
Probability O. 00f25 O. 175
I
(0.9)
I
Operates
-----' (0.1)
Release
L...--"""T""-----l
Ignition ~
Immediate
Ignition
Ignition
(0.5) (0.2)
Probability - 0.000 25
[= Explosion
Figure 1.5. Event tree for a LPG tanker accident (i.e., liquid propane gas).
Fault Trees
Fault trees are the best known and most widely used technique for developing failure
logic. The basic process adopted is to select an undesired "top-event" (e.g., leakage,
fire, or explosion), and trace it back to the possible causes which can be component
failures, human errors or any other pertinent events that can lead to the top event. This
procedure should be followed methodically, identifying the immediate precursors to the
top event, then the immediate precursors to these sub-events and so on.
TRANSPORTATION RISK - ENVIRONMENTAL AND SAFETY FACTORS 21
A fault tree is comprised of a complex of entities known as "gates" which, when the
logical combination of the input conditions is satisfied, produce a specified output
which is propagated. The construction of a fault tree follows a defined methodology,
but there is, as yet, no universally adopted set of symbols, which is unfortunate as fault
trees provide a vivid method of communication.
The fault tree is not a model of all possible system failures, or all possible causes for
system failure, as it only includes events which contribute to its top event. A fault tree
does not show sequences of events, although some complex gates can be "used to
illustrate conditions on the sequence of input events. This can be confusing and often
leads to difficulties in evaluating the tree. A fault tree essentially shows system states.
When compiled rigorously in "top-down" procedure, certain sub-events may appear
more than once in the fault tree. This is not anomalous and does not prevent the tree
being quantified. Attempts to draw the tree without repetition of events in the tree
usually lead to the "top-down" approach not being rigorously followed, increasing the
chance of omissions. An important feature that the analyst must consider is the
possibility of failure of independent items by a single cause.
Over the past few years a number of computer codes have been developed for fault
tree analysis. TABLE 1.4 lists some of the more readily available computer codes.
TABLE 1.4. Computer codes available for fault tree analysis
each of which can lead to fatalities within a different radius for each scenario. The road
can be divided into a number of segments, each of which will be related to a corre-
sponding population density. The following procedural steps are appropriate (Figure
1.6).
If one considers the movements of a road tanker carrying hazardous materials along a
route, for each sub-segment i of the route, there is a probability Pai for the tanker being
involved in a accident.
For each accident there is a number of possible accident scenarios Sj, each of which may
be considered to be fatal to individuals present within radius rj.
The number of people N present at the location of the accident, which may be affected,
depends on the density of the population D,
N=1CrfDj
Thus if one considers the passage of a tanker on route segment i, the probability of
someone being killed for scenario j is given by:
Pfataljty = Pai l'.j1C rf D j
The probability of someone being killed, due to the passage of the tanker on the sub-
segment i, is the sum of the probabilities for all possible accident scenarios:
Pfatality = L j P,.;l)Sj1C rf Dj
Fatality Probability = PaiDj L 1C rf l'.j
For anyone type of load, the term 1C rf l'.j is a constant which is independent of the route.
This term can be called the Severity Index, Sf, for the load in question.
Thus the probable number of fatalities from the passage of a tanker carrying load L along
sub-segment i is:
PaiDj(SI)L
and for the entire length of the route, it is equal to (SI)L LPaiDj . For anyone given
load, it is possible to compare the relative safety of two alternative routes by comparing
the term LP,.jDj , i.e., the population density, D;, along the route times the probability,
Pai, of an accident.
Cause-Consequence Analysis
Cause-consequence analysis is a technique which combines the ability of fault trees to
show the way various factors may combine to cause a hazardous event with the ability
of event trees to show the possible outcomes. Sequences and therefore time delays can
be illustrated in the consequence part of the diagram. The technique has considerable
potential for illustrating the relationships from initiating events through to end
outcomes. It can be ·used fairly directly for quantification, but the diagrams can become
extremely cumbersome. Because of this, cause-consequence analysis is not as widely
used as the first two techniques for complex systems, possibly because fault and event
TRANSPORTATION RISK - ENVIRONMENTAL AND SAFETY FACTORS 23
trees are easier to follow and so tend to be preferred for presentation of the separate
parts of the analysis.
Figure 1.7. Procedure for calculating the population risk of transportation routes.
Subjective routing factors in the selection (or elimination) of routes for the transport of
hazardous materials usually include:
• The location along the roadway (or in its vicinity) of sensitive land uses such as
hospitals, schools, old age person housing, churches or items of heritage or cultural
significance; or the location of sensitive ecological systems and natural landscape such
as park reservations, wetlands.
• Emergency and evacuation planning and infrastructure, including: the availability
of formalized emergency and evacuation procedures and plans; the location of emer-
gency response teams and their ability to respond to hazardous material release; access
and ease of emergency evacuation.
Subjective factors should reflect community priorities and values and should
preferably arrived at through community discussion and consensus. These factors are
particularly relevant in the assessment process when not one alternative is clearly
superior to the others. As such, whether or not the analyst chooses to select and apply
subjective factors which will depend upon the outcome of the risk calculations and how
conclusive the findings are.
24 CHAPTER 1
The guidance work sheets (Figure 1.8 and Figure 1.9) may be used as a guide in the
computation of the land use safety factors for assessment purposes.
(Explain): ...........................................................................................................................
Figure I.B. Example of a work sheet for compiling route characteristivcs and mandatory factors.
Figure 1.9. Example of a work sheets for computing land use safety factors
(Note: PHMA means the probability of a hazardous material accident).
TRANSPORTATION RISK - ENVIRONMENTAL AND SAFETY FACTORS 25
The work sheets (Figure 1.8 and Figure 1.9) summarize the factors of influence that
must be considered when dealing with regional risk assessment and safety management
related to the transportation of dangerous goods.
1.2. Traffic and Operational Factors for Road and Ran Transport
The following traffic factors influencing the transportation of dangerous goods reflect
the ability of a route to effectively and safely move the traffic flows using it:
The composition of vehicles by size and type is required to assess the road structural
adequacy as well as its operating level of service.
Traffic volume and composition along various sections and segments of the road
network may be obtained from published statistical information but preferably through
field screening surveys. Traffic volume may be expressed in terms of: annual average
daily traffic (AADT), and hourly traffic volume (average and peak). The directional
distribution of traffic should also be obtained. This information together with hourly
intersection counts can be used to estimate the peak directional hourly volumes along
all road sections within the study area.
Classification counts to establish the type of vehicles would differentiate, as a guide,
between: Light vehicles and Heavy vehicles (both rigid tankers and articulated tankers).
The level of service for a road section indicates the capability of roads for moving the
type and volume of traffic using it.
It describes these conditions in term of several factors such as speed and travel time,
traffic interruptions, safety, driving comfort. A possible designation of the level of
service is from A to F, with level of service A representing the best operating conditions
(i.e., free flow), and F level of service the worst (Le., forced or break down flow).
26 CHAPTER 1
TABLE 1.5. Example of one-way traffic volumes (PCU) for urban roads at different level of service
(for interrupted flow conditions)
TABLE 1.5 above shows suggested one-way hourly volumes for interrupted traffic
flow at different level of service.
A "road service volume" is defined as the maximum hourly rate at which vehicles can
reasonably be expected to traverse a point or uniform section of a lane or roadway during
a given time period under the prevailing roadway, traffic and control conditions while
maintaining a designated level of service.
The structural and geometric adequacy of the routes under consideration (to cater for
heavy vehicles carrying hazardous material) should be assessed. Routes with good
geometry (e.g., wider carriage-way with minimum horizontal and vertical curves) and
good line of sight should be selected in preference to routes of lesser quality. In
TRANSPORTATION RISK - ENVIRONMENTAL AND SAFETY FACTORS 27
situation, where for other reasons, a route in the latter category was selected then it
should be upgraded to provide better geometry and reconstruct, if necessary, the
pavement to cater for increased volumes of heavy vehicles.
The number of traffic signals is often used as a measure of delays along a route section.
A route with smaller number of signals would most likely be chosen as it would have
the potential for less delays.
Travel time for vehicles using a route indicate the congestion points as well as reflect
the level of congestion. Travel time information are usually available from transport
authorities or may have to be undertaken by way of field surveys. NAASRA has
suggested the following average vehicle travel speed for different level of service, as
shown below:
In case of an emergency which would require the closure of a route designated for the
transport of hazardous material, an alternative route should be available.
TABLE 1.6. Example of travel speed and flows for different categories of level of
service
(1.6)
where
PRj = Probability of a toxic release for route segment i due to a truck accident
TARi = Truck accident rate (accidents per vhc-km) for route segment i
P(RlA)i = Probability of toxic release due to a truck accident
Li = Length (kIn) of route segment i
The truck accident rate data used as default values in risk assessment and routing
studies should reflect the influence of highway geometry and traffic variables that have
a demonstrated relationship to truck safety. Two relevant variables are the highway type
(two-lane highway, freeway, etc.) and area type (urban/rural). It would be desirable to
also reflect the influence of other geometric factors such as lane width, shoulder width,
curves, grades, and intersections. In addition to truck accident rates, the distribution of
truck accident types also varies with highway and area type:
• Rural highways and urban freeways tend to have a larger proportion of single-vehicle
non-collision accidents, while
• Lower-speed urban highways tend to have a higher proportion of multiple-vehicle
collisions.
In general the probability of release given an accident involving trucks carrying
dangerous goods is much higher in single-vehicle non-collision accidents than in single-
or multiple-vehicle collision accidents. Thus the probability of a release given an
accident is also expected to vary for different highway and area types. Each accident-
involving vehicles should be treated as a separate observation (Le., an accident
involving two trucks should be counted as two accident involvement.
The data needed for analysis are:
Highway / Road Geometric Data. They should include: number of lanes, lane sepa-
ration, one-way/two-way, urban/rural. In principle different categories can be formed:
Highways
• multi-lane divided highways
• three-lane undivided
National and DepartmentaVCantonal Roads
• two-lane undivided (Departmental, and Cantonal Roads), with allowed
overtaking
TRANSPORTATION RISK - ENVIRONMENTAL AND SAFETY FAcroRS 29
A. Single-vehide Accidents
B. Multiple-vehide Accidents
heating oil, other toxic liquids or solid products) according to the official dangerous
products list Tank trucks may be full or empty at the moment of an accident. Further-
more the report should also contain an indication whether the hazardous cargo was
released or not, and the kind of consequences resulting from the accident. In cases
resulting in pollution of the environment, as a consequence of leak, fire, or explosion,
the appreciation of the danger to the environment should include a statement concerning
the kind of pollution (i.e., air, soil, and/or water pollution), and extent of contaminated
area.
Truck Volume and Accident Data. Traffic volume data: These usually include data on
Annual Average Daily Traffic (ADDT) and may also include either the average daily
truck volume, or the percentage of trucks in the traffic stream.
Average Truck Accident Rates: In selecting accidents for inclusion in the analysis, it
is important to use the same definition of a truck (type) that was used in obtaining the
truck volume. Since detailed data for Switzerland are not always available it will be
necessary to use accident data concerning the whole truck volume in the country
(indifferent of their type) or to use data for the totality of the motorized traffic. The
Swiss Federal Office for Environmental Protection, Forests and Landscape (BUWAL)!
has issued a "Handbook No. III to the incident regulations for assessing the risks of
traffic ways", that complementS the regUlation and ordinance on incidents.!7 In fact the
Swiss authorities are recommending their own method for estimating truck accident
rates. The Swiss methodology (i.e., BUWAL "Handbook No. nr') has been summa-
rized and will be presented in the chapters on rail and road).
The average truck accident rate for each highway/road class can be computed as the
ratio of total truck accidents to the total vehicle-kilometers truck travel for that
highway/road class. The corresponding formula is given as:
A ..
TAR. = __1)_
(1.7)
J ATM;
where
TARj = Average truck accident rate for highway class j
AU = Number of truck accidents in one year on route segment i for
highway/road class j
ATM; = Annual vehicle mileage (kilometers) of truck travel on road i.
Accident Severity
The . data should be compiled to show the truck accident severity distributions by
highway/road/street class. The percentage of fatal and injury accidents should be
indicated.
particular highway class is computed as the sum for all accident types of the proportion
of each type of accident, times the probability of release given an accident falling into
the category in question. The probability of release is the highest for the case of a
collision with a train, in case of overturn and run-off-road.
U.S Study of Accident Types. Another US study investigated whether the type of
accidents caused by trucking firms that haul hazardous materials differ significantly
from firms who do not haul these goods. The data were obtained from the US Federal
Highway Administration and covers the period 1986-1991. The raw data (13,498
reported accidents) were interpreted using a Poisson formulatIon in which the number
of accidents is the dependent variable, the explanatory variables are multipliable and
one takes the exponent of a coefficient in order to interpret it. The Poisson regression is
by definition non linear and fits an exponential curve to data. The method offers distinct
advantages when applied with count data (accident). 19, 20, 21, 22 For this study, a
regression with a constant and the log of the fleet kilometers was used as appropriate
relationship. A major advantage of the Poisson regression is that it can deal effectively
with datasets where a large proportion of the observations on the dependent variable,
i.e., accidents, take the value zero.
Experience with accident reporting has also shown that there are serious
inconsistencies and under-reporting of damage-only accidents. Hazardous materials
carriers in the U.S. are five time larger than carriers of non-hazardous goods in terms of
an annual fleet kilometers. In essence it was found that among hazardous materials
carriers, accident rates decline with firm size, but firms carrying hazardous materials
exclusively have an accident rate 11 % higher than the other firms and a rate of fatalities
and serious injuries that is 22% higher. Firms that carry hazardous materials in
combination with general freight, have an accident rate that is 18% higher and a
fatality/injury rate that is 24% higher. Private carrier are safer than for-hire carriers.
Haulers of gases in packages and liquids in tanks have the highest accident rates.
Carriers of hazardous wastes have the lowest accident rates. Long distance operations
32 CHAPTER I
(more than 160 km) are associated with accident rates 22% higher and fatalities and
injuries rates 53% higher than that of firms that are exclusively involved in short
distance operations. The accident on long-distance trip tend to be more serious and
result in a higher rate of accidents that involve fatalities and injuries, as well as more
property damage.
Vehicle Flow
(amount of traffic exposure)
Accident occurrence
..
(probability)
..
(fire likelihood)
Tank Rupture
(timing I rupture preuure)
Overpresaure Propagation
..
(pressure decay)
Damage Occ\IITeDCC
..
(damage threaholds)
Damage Evaluation
(monetary damage cost)
hospices of the local government. A similar approach can be applied to the transport-
ation. storage and handling of dangerous goods.
An important criterion in the assessment and selection of a route network for the
transportation of hazardous material is the relative cost of delays and travel time. The
analysis of this information would enable the determination of the economic
implications of particular routes for the transport of hazardous material and the
transport operator's requirements for practical transport economies.
Transport costs: They fail into two basis categories. i.e.• fixed costs and variable
costs (usually referred to as operating costs). Generally, the former costs do not vary
significantly with the vehicle-kilometers traveled. If the tanker carrying hazardous
materials need to change to a route of a longer or shorter distance. (only) the operating
costs will vary and be higher or lower respectively. In many cases, both operators' cost
requirements and operators' convenience result in the use of the shorter route
irrespective of safety implications.
Operating Costs. They are based on two main components: - a variable cost for
operating the road tanker (RT), and the cost of the driver's time (Dn.
where
RT : Unit cost component for the road-tanker ($Ikm; SFrlkm)
D : Distance traveled (km)
DT : Unit cost component for the driver's time ($Ihr; SFrlhr)
T : Time taken to travel the distance (hr)
TRANSPORTATION RISK - ENVIRONMENTAL AND SAFETY FACI'ORS 35
These factors could be reflected by the distance traveled and the travel time along
the route.
For the above, it is indicated that the main cost criteria when assessing or comparing
alternative routes for the transportation of hazardous materials is the expected increase
or decreases in travel time (the main component that influence operating costs).
An increase or decrease in operating costs of over 10% is considered to have a
significant effect on the cost of transporting hazardous materials. It is also considered
that the distance cost could increase further as long as the travel time was within the
10%. For example, a longer distance route could have, or be developed with less con-
gestion thereby resulting in a travel time about the same as the shorter route.
Summary
This chapter focused on the analysis and assessment of transportation routes for the
carriage of hazardous materials. It introduces issues related to the integrated risk
assessment approach to the safety transportation of hazardous substances, namely:
transportation risk and environmental land use factors, capability of the existing road
network and cumulative traffic implications and economic distribution considerations
and operator's requirements for practical transportation economics. The nature of the
major risk to be assessed is included in this chapter. Systematic methodological issues
are outlined for the assessment of transport risk, and procedures for the development of
estimated truck accident rates, transport operational costs, etc., are proposed.
36 CHAPfER 1
References (Chapter 1)
There are new trends in the chemical industry to manufacture dangerous products at the
site of utilization or, when possible, to modify the synthesis method in such a way as to
avoid the use of hazardous substances. Integrated concepts are developed, that provide
for the minimization of hazardous wastes through reduction, reuse, recycling and
recovery at source, as well as addressing treatment and disposal of material not
managed at the source. Historically, the siting of hazardous waste treatment and
disposal facilities has been and is still a difficult undertaking, raising public fear. The
transport and distribution of hazardous substances, such as petroleum products,
liquefied petroleum gases, ammonia, chlorine gas, pesticides, chemicalslpetrochemicals
and radioactive materials, inevitably involve the potential for incidents and accidents
which 11Uly result in death or injury to people, property damage or da11Ulge to the bio-
physical environment through the effects of fire, explosion or toxicity. An increasing
number of transportation accidents involving hazardous substances have occurred
worldwide. Such accidents with their resultant effects on people and the environment
have increased awareness in government, industry and the community and resulted to
think again of the risk assessment process applied to transportation of hazardous
substances. In that context, it is now recognized that the safety planning of transport-
ation routes, accounting for the type and nature of surrounding land uses, is an integral
component of the safety 11Ulnagement of hazardous substances transportation.
Delineating hazardous substances transportation routes is, as such, a significant and
essential complementary measure to technical and operational safety and environmental
controls on the hazardous substances containers and associated regulatory processes. It
is relevant to note that fixed installations are more amenable to local, organizational and
operational hazard controls. Transportation systems are dynamic systems with
additional external variables (e.g., drivers, traffic conditions, etc.) difficult to bring into
one overall control system.
The scope of regional risk assessment concerning transportation of dangerous goods
has to include all transport systems, such as rail, road, ship, and pipeline networks. The
main goal of such studies is to provide the (Federal, Departmental, or Cantonal)
Authorities with facts on actual risks incurred in a given region, on which to base
political decisions regarding the course and costs of mitigation actions and improve-
ments to be brought to existing transportation networks and distribution systems. One
has to learn from the mistakes done in the past and improved concepts and design
should be aimed at when planing new traffic ways and traffic control systems. Risk
cannot be eliminated totally, but reduction to an acceptable level is, most of the time,
38 CHAPfER2
possible. The cost of increased risk reduction have to be balanced with the social
benefits which can be expected from such reduction measures, since fmally the extra
expenses must be financed by both industry and Community. Quite often, in recession
crisis, abatement measures for environmental protection (noise, pollution) or for
improving the security of transportation systems and distribution networks are
postponed, because public money is simply not available at the moment.
Although the Authorities are aware of the danger and drawbacks of actual or planed
transport systems, it is quite often not possible to chose a different concept or another
layout due to natural constraints such as the presence of rivers, ground water, valleys
and mountains. In many cities built during the middle age, and which escaped bombing
during the second World War, the traffic conditions are difficult due to tortuous or
narrow roads. In many cases the whole traffic has to pass through the center of the city
or village and by-pass road are non-existent. What concerns the implementation of new
industries or new residential areas, the case is different since alternative choices can be
found and it is usually possible to minimize risk by sound selection of sites, avoiding a
cumulating of activities presenting a risk to the population.
The orderly development of industry and the protection of community safety
necessitate the assessment of hazards and risks. Recently the Swiss Federal Office for
Environment, Forestry and Landscape (BUW AL)1 has issued different handbooks No. I
through III, in the form of guidelines, of which "Handbook No. m, was prepared for
"assessing the risks of traffic ways". It is a complement to the Ordinance of April 1,
1991 concerning the "Protection Against Incidents" (Storfallverordnung, StFV)2.
The assessment of risk necessitates the establishment of criteria against which
judgments can be made as to the compatibility of various land uses. To date, risk
assessment criteria in Switzerland, have been applied on a case-by-case basis by the
different cantonal departments (e.g., Dept. of Planning, of Road Construction, etc.). In
Switzerland the departments in charge of "regional planning" and "risk assessment"
should now consider it appropriate to agree upon and formally issue a set of criteria,
based on accepted international experience, which shall assist in a consistent approach
to regional risk assessment and in the decision making process in the Community.
At this place it is worth noting that the denomination for the same kind of department,
or cantonal office/authorities may differ quite a lot from one Swiss Canton to another,
depending also of the respective official language in use (i.e., French, German, Italian,
and Romansch).
2.2.1. QUANTITATIVERISK.CRITERIA
The results of a hazard assessment is a quantification of risk. The level of risk which is
acceptable is a difficult question to answer. Quantitative risk criteria must be developed
by the competent authorities, which can be used by industry to provide a basis for
ASSESSMENT OF TRANSPOTATION RISKS IN A COMPLEX AREA 39
responsible decision-making on risks. Some of the problems which arise in this area are
those of:
• individual and average risk
• death versus injury
• maximum risk to employees
• maximum risk to public
• value of human life
• methods and criteria for comparing human loss and property damage
• multiple fatalities
• engineering feasibility
• allocation of resources, etc.
Several indices of industrial injury risk are used in the literature:
• Lost Time Accident Rate (number of accident per 108 hours).
• Fatal Accident Frequency Rate (FAFR), defmed as number of deaths per lO8
hours. (The Fatal Accident Rate (FAR) of the chemical industry may be
expressed as deaths/l08exposed hours).
• Death Rate per Annum (DRA).
One method of assessing the value of human life is to calculate the value of the
future production or services which an individual may be expected to give to the
community. A variant of this is to estimate future earnings, although these are not
synonymous with the value of the work done.
Risk C, fteria
The primary criterion is that the maximum risk to an employee is limited to the Fatal
Accident Rate (FAR) of the chemical industry (Kletz 3, 1971; Kletz4 , 1972). Another
primary criterion is that the maximum risk to a member of the public is limited to a
level comparable with other involuntary risks (target: risk of death < 10-7 per year
averaged over the whole population). The secondary criterion is that the value of a life
method is used to assess additional expenditure aimed at eliminating other hazards.
The risk criteria just described need to be supplemented by an additional criterion
for multiple-casualty accidents. In such a case it is assumed that there should be no
special weighting designed to reduce the probability of death due to a multiple-casualty
accident as opposed to death due to an individual accident. This can be represented on
a log-log graph of accident frequency versus number of deaths, using a slope of -1.
A further criteria which needs to be satisfied includes limits on the total frequency f
of accidents and on the risk r to individuals_ Thus:
n
f = "2Ji
i=!
(2.1)
1
r=-Lxj; ,
II
(2.2)
N i=1
Risk is defined as the likelihood of any adverse outcome. The suggested risk criteria are
probabilistic in nature. That is, they account for both the involuntarily consequences
(effects) and likelihood probability of hazardous events.
All activities associated with the storage, transportation and handling of dangerous
goods have an associated level of risk. Risks can be assessed and managed, but it is not
possible to eliminate totally a risk unless the activity itself is eliminated. In many cases
this simply leads to risk transfer (for example from train to road, or vice versa) which is
an important concept in risk assessment and management. The criteria are therefore
based on the concept of a residual risk, the acceptability of which should be established
in relation to various land uses.
A rational decision making process requires the establishment of a consistent
framework with standards to express the desired level of safety. Probabilistic Safety
Criteria (pSC), which are quantitative expressions for the probability of occurrence of
an undesirable event within a given period of time, can play the role of such standards.
The purpose of this section is to provide a general guidance concerning the setting
and applications of such criteria.
Public health risk should be assessed separately from environmental risk. In terms of
health impacts, occupational and public risks should be treated separately. Two
categories of risk apply as a result of direct or indirect impacts:
ASSESSMENT OF TRANSPOTATION RISKS IN A COMPLEX AREA 41
individual risk), each possible incident outcome is considered in tum and its frequency
(F) and the numbers of people (N) that could be affected by it, are recorded as an F-N
pair. The calculation of the number of people affected includes an allowance for the
average geographical distribution of the population.
In addition to the risk to people and property, the impact assessment process for
potentially hazardous storage, or loading/unloading installations, or transport systems,
must also consider the risk from accidental releases to the biophysical environment. Fire
and explosion hazards are of less relevance to the environment, in comparison to the
effect these hazards may have on people. Acute and chronic toxicity impacts are those
which must be chiefly addressed. Generally, there is less concern over the effects on
individual plants or animals. The main concern of fire and explosions hazards is instead
with whole systems or populations.
44 CHAPTER 2
The assessment of the ultimate effects from toxic releases into the natural ecosystem
is difficult, particularly in the case of atypical accidental releases. Data are limited and
factors influencing the outcome variable and complex. There may be no immediate loss
of plants or animals or other observable effects from single releases, but there may be
cumulative and synergistic effects. It is therefore appropriate to ensure that a thorough
review of available data are undertaken and best available information used in the
assessment process.
In many cases, it may not be possible or practicable to establish the final impact of
any particular release. It may be appropriate in such circumstances to assess the
likelihood of identified concentrations of concern occurring in the air, water or soil.
Where such criteria are used the assessment should remain on the conservative side.
Because of the complexities of such assessment and case-to-case differences, it is
inappropriate to specify general criteria. The acceptability of the risk will ultimately
depend on the value of the potentially affected area or system to the local community
and wider society. For example, where a rare or endangered ecosystem or species is
involved, a much lower risk level is necessary than where the potentially exposed area
or system is degmded and/or common.
The basis for a quantification of the risk from a hazardous or industrial activity is a list
of hazardous events, or groupings of like events which can be considered to produce
similar consequences. The frequencies of these events may be estimated. There may be
a range of possible outcomes from each event, depending on the different circumstances
which may prevail: for example wind direction, weather category, location of people
etc. Each of these cj.~cumstances must be defined and a probability assigned to them.
The aggregation of frequency and consequence analysis can therefore be complex,
although it is conceptually simple, because all analyses follow essentially the same
procedure.
Damage-causing events must be related to the undesired initiating events: for
example, the various possible outcomes arising from a release of flammable material
may be modeled using an event tree. The conditional probability for factors such as
wind direction towards ignition sources and chance of ignition at each source can then
be used to produce a frequency for the damage-causing event from the frequency of the
initiating event.
The consequences of each damage-causing event must be assessed: The usual
approach is to define ranges to selected casualty probabilities from a combination of
effect and vulnerability models (Figure 2.2). These causality probabilities may be
selected apd limiting ranges to each value estimated - for example, the probability of
casualties occurring at various over pressures could be used in conjunction with an
explosion overpressure model to produce radii to selected casualty probabilities.
ASSESSMENT OF TRANSPOTATION RISKS IN A COMPLEX AREA 45
I
Assess the Consequences
Damage to people (probability of deaths, injury)
Damage to property (frrelstructural damage)
Damage to environment
Assess fmanciallosses and associated cost
(operating loss, remedial actions)
The selection of probabilities will usually depend on the available data underlying
the vulnerability model used - the analyst should be wary of using probit type relation-
ships to produce a large number of casualty probability bands as this not only compli-
cates the analysis, but the degree of detail would not be supported by the basic data.
46 CHAPI'ER2
Having obtained the frequency and casualty probabilities for each damage-causing
event under consideration, the risk relationships are derived in the following manner.
Taking each event in turn, the number of people present in the area covered by each
casualty probability band are multiplied by the appropriate casualty probability,
producing, by summation, the- total number of people predicted to be affected by each
event. The overall frequency-consequence relationship can then be drawn up from the
number (e.g., people, property, land, or water resources) affected and the frequency for
each event. Expressing the risk in terms of the frequency distribution of multiple
casualty events (N-F-curves) is known as "calculating the societal risk".
QUIlntijicalion of Risles
Step 1: Estimate the frequencies of these events.
~: Define the circumstances concerning a given scenario and assign a conditional
probability to each of them to produce a frequency for the damage-causing event from the
frequency of the initiating event.
S1HL1: Define, from a combination of effect and vulnerability models, several conse-
quence ranges which are related to the selected casualty probabilities.
Step 4: Derive the risk relationships for each event, and multiply the number of people
present in the area (corresponding to each casualty probability band) by the appropriate
casualty probability. Finally the total number of people affected by each potential event is
obtained by summation.
~: Calculate the societal risk by estimate the overall frequency-consequence relation-
ship from the number (e.g." people, property, land, or water resources) affected, and the
frequency for each event. Express the resulting risk in terms of the frequency distribution of
multiple casualty events.
Step 6: Calculate the individual risk at a given location by taking the casualty probability
at that location for each damage-causing event, and multiply it by the frequency of that
event. Obtain the individual risk from all such events, (and therefore from the activity as a
whale) by summation over all the events.
The individUill risk at a location is obtained by taking the casualty probability at that
location for each damage-causing event and multiplying it by the frequency of that
event. The individual risk from aU such events, and therefore from the activity as a
whole, is obtained by summation over an
the events.
The final expression of individual and societal risk then incorporates the likelihood
and severity of all the outcomes of the scenarios that have been considered for particular
location. The following sections illustrate the way in which the calculations can be
carried out (see also Figure 2.3).
Note that in above formula the sum is computed over all people within the area
affected by the hazardous event.
When evaluating the number of people affected by a given hazardous event, the
following expressions are often used:
When evaluating individual risks, assuming the "person at most risk approach" is
=
adopted for a given location, then it is usually assumed that P DeC 1. Expanding PIwz in
terms of the likelihood of a release Prol and the probability that it will give rise to the
hazardous events p.se leads for the individual most at risk to the expression:
(2.7)
ASSESSMENT OF TRANSPOTATION RISKS IN A COMPLEX AREA 49
wbere
PreI = probability of gaslliquid release
Puc = probability of the hazardous event occurring as the result of the
release
For a number of releases and hazardous events, individual risks are summed, over
all possible accident scenarios and types of accidents, i.e.,
(2.8)
The steps summarized in Figure 2.5 may be give guidance in the computation of
individual risk of fatality.
small but incremental increases in population over a period of time which may
eventually have significant implications on a cumulative basis.
In view of these considerations we suggest that:
• Judgments on societal risk shall be made on the basis of a qualitative approach
on the merit of each case rather than on specifically set numerical values.
• Individual fatality and injury risk contours at the individual risk criteria levels
(appliCable to the various land use categories) should be established. People
density within each risk band should then be established and the suitability of a
particular regional development judged on the basis of the location, density and
vulnerability of the exposed population.
In practice, the frequency of each potential accident and the number of people that
may be affected by each accident should be estimated wherever practicable. F-N curves
should be generated and used as an input to the risk analysis. An example of societal
risk curves for some human-caused events, that was adopted by the USA, is shown in
Figure 2.6.
In general, the individual risk criteria for particular land uses are set with regard to
the lower end of the range of typical population densities for that use. Where individual
developments (industrial or residential), or a number of developments taken together
result in substantially higher densities, their location relative to the suggested criteria
contour should be carefully considered. Generally the closer to the limit for that type of
land use, the lower the acceptable population density should be.
The analysis should also consider the specific vulnerability of such development and
popUlations, taking inte account the specific vulnerability of the people (age.ct, young,
disabled, etc.); the topography and access to and egress from the development and
locality for emergency response and evacuation; other emergency infrastructure and
design of developments (e.g., glass facades towards an explosion hazard, people in the
open air where the hazard is toxic gas).
10
-><
1\
10-2
~
c:
~
....... 1Q-3
is'
c:
CD
::l
~ 10-4
LL
1fr5~----~~~--~----~-----4----~
1fr6~----~----~~----+-----~----~
10~L-----~------~~~~------~----~
10 102 103 1Q4 10S 1()6
Fatalities (x)
Figure 2.6. Societal risk curves.
ASSESSMENT OF TRANSPOTATION RISKS IN A COMPLEX AREA 53
For the purpose of injury, a lower heat radiation level (relative to that level which
may cause fatality) is appropriate. A heat radiation level of 4.7 kW/m2 is considered
high enough to trigger the possibility of injury for people who are unable to be
evacuated or seek shelter. That level of heat radiation would cause injury after 30
54 CHAPTER 2
seconds exposure. Accordingly, a risk injury criteria of 50 in a million per year (50 x
10.6 per year) at the 4.7 kW/m2 heat flux is suggested for residential areas. International
experience with the implementation of that criteria indicates that it is achievable and
appropriate.
Explosion Effects
Overpressure
(kPa)
Using a similar analysis to that adopted in establishing a heat flux injury level, it can
be suggested that an explosion overpressure level of 7 kPa be the appropriate cut-off
level above which significant effects to people and property damage may occur.
Accordingly, an injury/damage risk criteria of 50 in a million (50 x 10-6 per year) at the
7 kPa explosion overpressure level is suggested for residential areas. International
experience with such implementation confirms this level as appropriate.
ASSESSMENT OF TRANSPOTATION RISKS IN A COMPLEX AREA 55
buildings and as such is appropriate as the basis for a risk of damage criteria A
probability of 50 in a million per year is also suggested for assessment purposes.
These criteria do not remove the need to consider higher consequence levels at
lower frequencies. The hazard analysis should consider the whole picture not just the
nominated quantitative criteria
The qualitative and quantitative results of the risk analysis can be applied in the
assessment process as follows:
a) Risk impacts at various distances from a planed or existing storage, respectively
loading/unloading installation, pipeline routes, or selected raiVroad routes may
be compared against safety targets or criteria A judgment can be made about the
hazard impact. A general principle of assessment is that the risk impacts from the
new development should be well below the levels of risk which people and the
environment are regularly exposed to from similar development and other
sources.
b) The analysis should particularly highlight the major contributors to the risk and
their nature and extent and, secondly, areas where risk could be eliminated, or
cost-effectively reduced. These results can be used to develop prevention and
protection measures including priority allocation of resources for hazard control.
The requirements for transportation of goods are changing due to new forms of
industrial production methods like JIT ("just in time"), minimal storage requirements
and trading. Transports of dangerous goods, especially by road, are recognized as a
problem by some members of communities and there are groups which consider it to be
a very risky activity, especially what concerns transportation of radioactive spent fuel
elements (that are, as every body knows, transported in very heavy fuel casks designed
to withstand a drop from several meters).
Accidents involving "thin walled" tanker releases are, in some countries, high enough
for direct estimation of the conditional release probability. Looking at "thick walled"
tankers, releases are scarce, and conditional release probabilities can be determined by
applying a deterministic failure analysis to a statistical description of accident
environment stresses (Clarke, et al}6 1976).
True Accident Rates. Recently methods have emerged for assessing the true accident
rate, i.e., the anticipated number of accidents after a certain number of accidents have
occurred over a given period. One of these methods is the ''Empirical Bayes Method"
and the other make use of Poisson distributions, which means that the expected values
for accidents in a population of intersections or road segments of a similar type are
gamma distributed. To use these methods, a number of accidents that is not too small
must have occurred for every type of accident under consideration. This again may not
be the case with road-tankers, for which very few accidents have been registered.
Methods for estimating accident frequencies and comparing alternative routes shall be
presented in the next chapters.
Swiss Accident Statistics. There is generally a problem with the availability and
appropriate reporting of statistical data on dangerous goods' accidents, this especially
in Switzerland, where for instance the use of computer in public administrations is quite
recent. The problem is even enhanced due to the Swiss federalist system according a
large independence to the Swiss Cantons, which all have a different way of reporting
accidents and preparing statistics. Mostly in the past, there was no preferential diffe-
rentiation between the different kinds of traffic accidents, and it is actually still difficult
to know the exact number of accidents involving road-tankers or truck transports of
other dangerous goods. Usually, accidents are registered and reported by the Municipal
or Cantonal Police (urban/highway traffic) using different protocols and computer
software/systems. There is no Federal Accident Database available for Switzerland.
Reporting of Accidents. Most of the time, in case of material damages involving cars
and trucks only, the participants do announce the damage to their respective insurance
and the police is not involved or informed at all about that accident. When people are
injured, the police must be called and will register the accident. As long no material has
been released to the environment, the police will usually not record whether the truck
was empty, or transporting earth, construction material or some dangerous goods, nor
record what kind of product was involved in the accident. There is also another problem
with historical accident data that is due to traffic intensity and road communications
changing over the years, and thus the statistical data may not always reflect actual
traffic conditions. Therefore accident rates cannot always be satisfactorily extrapolated
on the basis of historical data (Abkowitz,17 1988).
• selective object criterion (selected objects like ecological systems, etc., should
not be damaged or harmed by water soluble or other substances).
Dynamic rerouting would be an alternative for managing the risk of dangerous
goods transports. Efforts in this direction are in progress, especially in the United
States. The method presupposes the availability of dynamic traffic control systems
based on the use of Global Positioning System (GPS) located on each truck, which
supply the exact position of vehicles and a radio-communication system which transmit
the data to the traffic manager. Such solutions are costly and no legislation, at present,
can oblige any trucker company to equip its vehicles with such systems. The congestion
of traffic is so severe today that it would be useful to take corrective measures. One task
for a traffic control system will be to distribute the demand for road resources in time
and to optimize the traffic flow (private cars, trucks and public transports) and to select
appropriate roads dynamically, passing the information to the road-user equipped with
an automatic display, to redistribute the traffic by proposing alternative roads. Another
solution which has already been adopted in several countries consists in controlling the
traffic flow by modifying the allowable speed on highways to avoid traffic queues.
Irrespective of which risk criterion is selected, methods of implementation are
always needed so that the routing will have the desired effect of reducing transportation
risks. In addition to a method for determining which roads present a lower risk, a way
of informing the drivers in time (i.e., in advance) of the presence of obstacles, road
repairs, deviations, and traffic queues has to be implemented. At present the technical
installations and the necessary legislation are missing. Actually information for road-
users is broadcast on special radio frequencies. Some car-radios allow the current radio
emission to be automatically interrupted for receiving special traffic messages. The
main dr3w-back of this system is that information is usually not quite up-to-date or
reaches the driver much too late, when this one has already been stopped by a queue on
a highway.
often compared to rail transport on the basis of the number of fatalities per 1000 t, as
shown in the example of Figure 2.7.
Environmental data (weather categories, ecosystems, forests, urbanlland population)
and other roadway elements, such as the number of at grade road/rail crossings, number
of turns, type of road/highway, etc., have to be taken under considemtion as well and
risk factors and accident scenarios developed. The corresponding traffic, accident and
release data are compiled and corrected to reflect the conditions of selected road
segments. The population density along the transport corridors must also be estimated.
The dangerous goods are usually classified according to danger classes, the quantities
transported must be known or estimated and the transport risks must be calculated for
each danger class and each road or rail segment.
One as to differentiate between the risk of an accident and the risk of a spill, fire or
explosion. Risk mitigation measures and the overall system sensitivity to these
measures have to be evaluated. To establish optimum routing conditions, two types of
modeling conventions may be used:
• Minimum time travel, where trucks are routed on the basis of least time from one
point to another. This inherently minimizes risks as it reduces the time exposure
of a vehicle on a roadway.
• Minimum accident probability, where trucks are routed on the basis of least
accident probability from one point to another.
Depending on the region being assessed, alternative routing could be found possible
or the advantage of one transport medium against another become more apparent.
Redistribution of some traffic may be considered in order to reduce the risk to popu-
62 CHAYI'ER2
Large industrial areas accommodate an extensive number of risk sources and activities
of varying nature and extent. Such sources may include operating process plants,
storage terminal, transportation activities, etc. Ideally, a cumulative assessment of such
risk should include a detailed hazard analysis and quantified risk assessment for all
industrial facilities and associated activities. In many cases, however, because of limited
resources and time constraints, a preliminary evaluation of the various risks is needed,
in order to establish which activities should be the focus of detailed risk assessment in
priority to others and where ~sessment resources should be allocated in the framework
of regional risk assessment and land use planing. Method for classifying risks in a
region (which also can be applied to the risk of transportation, handling and storage of
dangerous products) have been proposed in an Inter-Agency program on the
"Assessment and Management of Health and Environmental Risks from Energy and
Other Complex Industrial Systems",19 The program in question was sponsored by
UNEP (United Nations Environment Programme); WHO (World Health Organization);
IAEA (lntematiomd Atomic Energy Agency), and UNIDO (United Nations Industrial
Development Organization).
(2.9)
• Step 1: For each dangerous good transported: select one route (road, railway, waterway,
pipeline); select segments of 1 kIn in length; choose the most hazardous segment out of
the list (high population density, high traffic load and bad traffic safety).
• Step 2: Select from the table the average probability number for each hazardous
substance (or group of substance) and for each road segment under consideration.
• Step 3: Estimate the probability number correction parameter 1\: taking into account the
safety conditions of the transport system.
• Step 4: Estimate the probability number correction parameter lit taking into account the
traffic density (number of transport units per year).
• Step 5: Estimate the probability number correction parameter IIp taking into account the
wind direction
• Step 6: Calculate the probability number Nt,s using the above formula.
• Step 7: Convert the probability number into a probability, P.
• Step 8: If a portion of a roadlraillwaterway/pipeline is exposed to a risk: of accident
from different hazardous substar.::es, the frequencies (accident frequency per kIn and per
year) calculated for each of them have to be grouped under the defmed classes of injuries.
• Step 9: Repeat all the steps for each road segment and each substance transported on
that road segment.
Figure 2.B. Procedure for estimating the probabilities of major accidents resulting from transportation
of hazardous materials.
The probabilities of major accidents for storage or fixed installations can be
estimated according to the formula:
(2.10)
Nj,s Average probability number for the installation i and the substance s
nf Probability number correction parameter for the frequency of loading!
unloading opemtions
nl Probability number correction parameter for the safety systems associated
with flammable substances
no Probability number correction parameter for the organization and
management safety
np Probability number correction parameter for wind direction towards the
populated area.
ASSESSMENT OF TRANSPOTATION RISKS IN A COMPLEX AREA 65
In order to calculate the frequency (Pi,., number of accidents per year) of accidents
during transportation, the so-called probability number should first be estimated:
(2.11)
The necessary data, tables and correction factors for this method can be found in the
afore-mentioned UN-Interagency report (1993).
The procedural steps for estimating the probabilities of major accidents resulting
from transportation of hazardous materials is given in Figure 2.B.
Fixed Installations. A similar approach is taken for fixed installations. The estimation
of the societal risk is done according to the procedure of Figure 2.9.
• Step A: Classify each activity using a scale of consequence classes (0-25, 26-50, S1-l(lO,
101-250, 2S1-5OO, >SOO fatalities/accidents) and a scale of probability classes (frequency =
10-3 to 10-7 accident per year)
• Step B: If a certain industrial activity presents risks to the population from different
substances which can cause accidents independently from each other, sum up the risk from
substances which have the same class of consequences.
• Step C: For risk classification, prepare a matrix of the activities showing frequency vs.
consequences.
Figu.re 2.9. Procedure for estimating the probabilities of major accidents resulting from fixed installations.
Hazardous materials are moved by the foUowing modes of transport: road, rail,
waterway, pipelines, ship, aircraft. The type of material transport used may greatly
affect the storage requirements. A mode of transport which involves large but
infrequent deliveries is in general likely to require more storage than one which gives
continuous flow. Actually the trend for delivering electrical or mechanical components
or chemical products to the industry is to use the approach called ''just in time delivery".
This way of distributing goods is, unfortunately, requiring a much larger number of
66 CHAPTER 2
Transport nwy present a hazard to: (a) the transporter. (b) the crew. (c) the public.
The relative importance of these varies between the different nwdes of transport.
This section reports some analyses of transport accidents, and describes the main
factors contributing to accidents.
Transportation H(J1Jlrtis
The hazard presented by the transport of chemicals are:
• Fire
• Explosion
• Toxic release of conventional toxic substances
• Toxic release of ultratoxic substances
Road accidents leading to the release of hazardous nwterials are due in most cases to a
tanker crashing or overturning.
Big road disasters are for instance: San Carlos de la Capita, Spain (1978), in which
a propylene tanker crashed into a holiday camp site, cost the lives of some 211 people.
The explosion at St. Amand-Ies Eaux, Frct.'lce (1973), and the toxic release on the
Southwest Freeway, Houston, Texas (1976), etc.
Rail accidents in which there is an escape of hazardous material are mostly due to
train crash, or derailment. Mechanical features of rolling stock and faults in the rail
track are frequent causes of these. Another hazard of railway is fires along the track
caused, in the past, by steam engine or break shoe sparks.
• Containment failure
• Accident impact
• Loading and unloading operations
It is emphasized that there are many factors which determine the failure rate of an
equipment and that the range of failure rates observed for an equipment can be quite
wide. Two prime sources of failure rate data for fixed installations are the pUblication of
the UKAEA (U.K. Atomic Energy Authority), including the SRS (System reliability
Service) and the Reactor Safety Study (1975) by the AEC (Atomic Energy Commis-
sion, USA).
68 CHAPTER 2
Spillage and Evaporation of LNG and Ammonia on Water. For a rapid spill of LNG on
water the evaporation time is short and a puff dispersion model is applicable. But for a
continuous spill the evaporation rate rapidly reaches the spill rate and this evaporation
rate then determines the pool diameter reached. The evaporation rate has been estimated
to be about 0.19 kg/m2 s, which corresponds to a liquid regression rate of 4.7xlO-4m/s.
Similarly, it is stated that for a rapid spill of ammonia on water the evaporation time is
short and that all the ammonia will evaporate, except for about 20% which dissolves in
water.
A cloud of cold LNG tends to behave like as a heavy gas cloud. The model of the
source which is assumed can have significant effects on the results. The release may be
continuous or instantaneous. The source may be of fixed size or increasing size, as with
LNG spilled on to water. The process of release mayor may not involve the
entrainment of large amounts of air. The description of the source elements must be
done precisely in order to model correctly the scenario.
ASSESSMENT OF TRANSPOTATION RISKS IN A COMPLEX AREA 69
Canny struly
Potentially haz/lrdous material in the Canvey Island IThurrock areas, their location and the
quantities stored and in process were identified. The investigation identified several
principal haz/lrds in the area. These are:
I) oil spillage over bund
2) UfG vapor cloud release (l,OOO t)
3) [PG vapor cloud release (1,000 t)
4) ammonium nitrate explosion (4,500 t of 92% solution)
5) ammonia vapor cloud release (1,000 t)
6) hydrogen fluoride cloud release (100 t)
The following cases were considered in the Canvey study:
• a severe fire might occur if there is an escape offlammable liquids from storage.
• a severe vapor cloud fire and/or explosion might occur if there is a spillage of LNG or
LPG so that a vapor cloud forms and ignites.
• a severe explosion might occur if there is a rupture of an ammonium nitra~1! storage tank.
• a severe toxic release might occur if there is a spillage of ammonia or a rupture of
storage or process plant containing hydrogen fluoride.
Canvey Study: Probabilities of Various Occurrences and oftlwir Consequences
The frequencies indicated below are for:
• Spontaneous failure ofpressure vessel (LPG, ammonia, HF): 1(J5_10-4/year
• Failure of LPG pipework (whole refinery installation): 5 x 1(J3/year
• Large vapor release (LPG filling point): 5 x 10·3/year
• Serious fatigue failure for a UfG tank (above ground): 2 x 1rr/year
• Overpressurization by overfilling for a LNG tank (above ground): 10·s _10-4/year
• Rollover involving structural damage for a LNG tank (above ground): 1(Js -1 rr /year
• Unconfined vapor cloud explosion in a refinery: 10.3/year
• Major fire in refinery: O.I/year
• Failure of butane pipeline (15-20 cm dia.): 3 x lO-4/km.year
• Derailment of rail tank car: 1 x 1O.fi!train km traveled. Probability of overturning given
a derailment: 0.2
• Frequency of accident of road tanker involving spillage: 1.6 x 10-8/Ian traveled
improvements during the last decade. Assessment of the comparative safety of the
different modes of transport may consider only the risk to the public or they may take
into account also the risk to terminal operators and to transport crews such as drivers.
The latter may be injured or killed in road accidents and the inclusion of such accidents
make a significant difference to safety calculations for road transport. There is however
no transport crew for a pipeline.
the existing analyses of the state databases and to apply the result to the specific routes
of interest.
Train accident rates are more straightforward than those for trucks because there is
basically one source of accident data for a train and one source of exposure data. In the
USA, there exists a Railroad Accidentllncident Reporting System (RAIRS), which is a
database of accident reports (Abkowitz and List30 , 1986) submitted by the railroads to
the Federal Railroad Administration (FRA). The Association of American Railroads
(AAR) also maintains a database (TRAIN II) on railcar movements. The FRA Office of
Safety (OS) publishes an annual accident/incident report, entitled Accidentllncident and
Inventory Bulletin.
The AAR, the Railway Progress Institute (RPI), and the Chemical Manufacturers
Association (CMA), in the United States, are jointly developing a hazardous materials
transportation risk analysis mode1. 31 The accident cause portion of the model will
include the effects of accident type (collision, or non-collision), train speed, track class,
railroad type, train length, car placement within the train, tank car type, and car
protective features.
In the case of railroad transport, many countries have negligible history of serious
accidents to liquefied gas rail tank cars, and reliable statistics directly assessed from
these are usually not available. However one may find a large amount of data on
collisions and derailments and the type of location. It is from this information that a
series of rail tank car failures rates has to be assessed for a variety of locations.
The four principal modes for bulk transport are highway, railroad, pipeline, and
waterborne. In the case of highway transport, single loads up to 10-25 tons move under
various constraints of time or route. With railroad transport, the routes are more
restricted, and many other operating constraints can exist. Unit loads can reach up 100
tons and, in Europe, full train loads of up to 1500 tons can be carried. In North
America, mixed train loads with several thousand tons of hazardous chemicals are not
uncommon. Water transport varies from small unit loads, often on barges moving in
canal systems, up to bulk carriers moving loads in one or more large refrigerated tanks
over the seas.
In the first three modes of transportation, the presence at any location of the
hazardous material in tanks is usually of a transient nature, as the vessel or vehicle
passes. For most of the time it will not be present. There are some locations, however,
where it may be present for extended periods of time, such as during loading or
unloading operations, or when it is stored in large storage yards, locks, or vehicle parks.
In contrast, the material will always be present in a transport pipeline at location
through which it passes. In assessing the failure frequencies at each location, due
account of these factors must be taken.
72 CHAPTER 2
Although data on US rail incidents are easily available, it was felt the differences
between the design standards and operating practice made this data inapplicable to the
British situation. However US road dnta could be used, and by appropriate modification
to exclude those events which could not or were unlikely to occur in Great Britain, spill
frequencies were derived. Fault tree analysis was used to develop the possible cause of
events which could lead to ~uipment leaks. These were then used to derive approp;·iate
equipment spill frequencies for both rail and road transport of LPG, ammonia and
chlorine. For flammable spill events it is also necessary to consider the probability that
spill will then be ignited and whether this will take place initially, or at some later time
once a flammable cloud has developed. In the paper of Purdy,32 tables of frequencies of
spill against cause, substance, and transport mode, as well as ignition probabilities for
flammable substances are given for conditions prevailing in Great Britain.
Regarding consequence analysis, there are several constraints which act to limit the
range of possible events:
• Simultaneous loss of content for multi-compartment tankers is extremely unlikely.
• Small release of flammable material are unlikely to ignite.
• Above a certain hole size, either the release of pressurize, liquefied gas will be so
rapid that it can be considered equivalent to an instantaneous release on vessel rupture,
or the hole will lead to a propagating failure of the pressure vessel.
• In the rail environment ignited jets of LPG are unlikely to create significant hazards,
unless they impinge on other LPG tankers which then BLEVE. The BLEVE frequency
used should normally include such a cause.
ASSESSMENT OF TRANSPOTATION RISKS IN A COMPLEX AREA 73
• It was assumed that 10% of the release from pressure vessels were instantaneous and
could be modeled as the entire loss of contents.
• In the case of toxic materials the cloud was assumed to contain 100% of the tanker
contents
• For LPG, twice the adiabatic flash fraction was assumed to enter the vapor cloud.
• Containment effects of roads and drains can significantly affect the shape and
dimensions of the hazard zone (especially for flammable liquids).
For motor spirit two cases were considered:
a) tanker remaining on rail or road (confined by kerbs, drains, etc.), and
b) tanker left the road or rail track (unconfined case to be modeled as a circular pool).
Figure 2.13. Failure cases used for the analysis of rail and road transport risks in Great Britain.
Spreading pool expressions can be coupled with a "drain model" and a ''fire model"
to estimate the maximum area affected:
• The immediate ignition case were calculated assuming that the release occurred
over a finite time and that the pool size was the maximum possible after
regression (regression rate is equal to the spill rate).
• In the delayed case, the pool was allowed to spread to its maximum before
iL~tion took place.
• The possibility of soft BLEVE fire ball of a gasoline tanker does not seem likely, since
the it will at least take an hour of prolonged heating before a rupture occurs with only
two relief valves (out of four) working. For LPG the type and extent of hazard depends
on the mode of release, and whether and when it is ignited. BLEVE are much less likely
on the road.
• Vapor Cloud Explosions (VCE) are very unlikely on both road and rail. The risk and
consequences can be estimated using a simple consequence model such as TNT.
Outside the flammable cloud, the probability of death due to the effects of overpressure
from a VCE is low.
• The risk from released toxic gases, such as chlorine and ammonia, is very dependent
on the accuracy of the dispersion modeling. When calculating the societal risk, the
crosswind extent of the cloud is as important as the downwind hazard range.
In 1985, the Health and Safety Commission (HSC), in Great Britain, has examined
the potential risks of the transport of certain dangerous substances and the results were
presented in a HSC report34 in 1991. The terms of reference covered not only the
consideration of major hazard aspects of the transport of dangerous substances but also
the identification of appropriate control measures and advice on additional action that
74 CHAPfER2
might be necessary. They did not include radioactive substances, transport by air or by
pipelines or risk to the environment.
Transport in Bulk, of Toxic and Flammable Substances by Road and RaU in Great
Britain
It is concluded, that for Great Britain, the safe routing of materials with large hazard
ranges may be more easily achieved by road. While the natural separation afforded by the
rail system may make this mode more suitable for lower hazard materials.
It was recognized that existing accident statistics alone provide an inadequate basis
on which effectively to judge the risks involved in rare but potentially high
consequence events. It was concluded that a more appropriate approach is to use the
quantified risk assessment (QRA) methodology. The report acknowledges that the risks
estimated by this method are subject to some uncertainty, but judges that QRA has
provided the best estimates of the risks involved, provided valuable insights and
assisted the Sub-Committee in reaching its conclusions.
QRA can only provide numerical estimates of the risks; it cannot, by itself, provide
conclusions as to whether or not those risks are low enough or too high. Such judgment
are reserved to the decision-maker.
It is essential that chemicals which are to be transported are adequately packaged and
labeled. The packaging and labeling of chemicals is the subject of BEC Directive
67/5841EEC, as amended by subsequent directives. The provision of these directives are
implemented in the Packaging and Labeling of Dangerous Substances Regulations
1978. Further recommendations on the marking and labeling of dangerous chemicals
are given in "Marking Containers of Hazardous Chemicals and Labeling Chemicals", A
76 CHAPTER 2
Guide to the Regulations for the Labeling of Hazardous Chemicals by the CIA (1970
Item 13), 1977 Item 14).
Some aspects of packaging have been covered by Kieselbach in his report. 39 The
various aspects of packaging and labeling of chemicals for road transport include:
(1) packaging and labeling of the chemical, (2) loading and securing of the container
on the vehicle, and (3) labeling and documentation of the vehicle. Loads should be well
secured to the vehicle. The Code of Practice - Safety of Loads on Vehicles by the DOE
(Department of Environment) covers this aspect.
It should be noted that in Switzerland there does not exist a Ministry of Traffic. The
responsibility for ensuring the correct application of the regulations is imparted to the
Cantons and to the Swiss Confederation.
Road Transport
• ADR ordinance, "Accord Europeen Relatif au Transport des Marchandises Dangereuses
par Route (ADR)",35 or "Europiiisches Uebereinkommen uber die internationale Beforde-
rung gefiihrlicher Guter auf der Strasse (ADR)", vom 30. September 1957 (status by: 1. Mai
1985)., and by its Swiss equivalent below.
• SDRIRSD ordinance: "SDR, Ordonnance Suisse Relatiw' au Transport des Marchandises
Dangereuses par Route, 1972)". This ordinance has beenfdly revised in 1985 antifollows
now the description of the international ADR agreement. The German terminology is now:
"RSD, Beforderung gefiihrlicher Guter auf der Strasse". Verordnung jUr die Beforderung
gefiihrlicher Guter auf der Strasse (RSD) vom 17. April 1985.
Figure 2.15. Other Swiss regulations and Federal laws concerning the transportation of dangerous goods
by road..
Figure 2.17. Other Swiss regulations concerning the transportation of dangerous goods by rail.
The Swiss ordinances and other Swiss regulations can be obtained from the Swiss
Federal Printing and Materials' Office, EDMZ ("Eidgenossische Drucksachen und Ma··
terialzentrale") in Bern, Switzerl~d, .
Since 1.1.1992, the Swiss Federal Dangerous Goods Inspectorate (EGI, "Eidgenos-
sische Gefahrgut-Inspektorat") of the Swiss Association for the Control of Pressure
Vessels, SVDB ("Schweizerisches Verein fUr Druckbehlilteriiberwachung") is now res-
ponsible for the examination and licensing qf transport containers.
The carriage of hazardous material by air is the concern of the Restricted Anicles Board
of the International Air Transpon Association (lATA). The lATA Restricted Anicles
Regulations place restrictions on the type and quantity of chemicals which are carried
and specify requirements for packaging and labeling. These regulations are, also in
Switzerland,40binding on the lATA member airlines (Figure 2.18).
The regulations list some 2300 substances requiring special packing and handling.
These include flammable liquids and solids, combustible liquids, explosives, oxidizing
materials, organic peroxides, corrosive materials, poisons, noxious or irritating sub-
stances and radioactive materials. A distinction is drawn between passenger and cargo
aircraft. The carriage of radioactive substances is governed by the lATA Regulations on
78 CHAPTER 2
Radioactive Materials which embody the IAEA Regulations on the Safe Transport of
Radioactive Materials.
• Verordnung uber die Beforderung gefiihrlicher Guter mit LuftJahrzeugen (SR 748.411)
vom20.9.1984
The Swiss Federal Office for Civil Aeronautics (BAZL)i is responsible for the
application of the regulations concerning air transportation of dangerous goods. This
instance, however, does not have its own experts for the domain concerning the
transportation of dangerous goods.
Figure 2.20. Swiss Federal decree concerning the international SOLAS agreement.
This means that the IMoo code of the IMO is applicable to Switzerland which has to
fully comply to it. The Swiss Federal Ministry responsible for the application of this
regulation is the "Schweizerisches Seeschiffahrtsamt Basle".
• Verordnung iiber die Beforderung gefdhrlicher Guter auf dem Rhein (ADNR) yom
15.7.1987 (SR 747.201).
Figure 2.21. Swiss ordinance concerning the transportation of hazardous materials on the Rhine.
Since 1.1.1993, there are several changes, modifications and improvements which were
brought to the different ordinances. Detailed information concerning such changes can
be found in Kieselbach 39 (1993).
TABLE 2.4. Swiss Authorities responsible for the application of the ordinances and regulations
The Swiss Authorities mentioned in TABLE 2.4 are responsible of the application
of the different regulations and ordinances.
via), and of other institutions such as the nuclear research institute PSI (paul Scherrer
Institute) at Villigen, the AC laboratory of the Swiss army (Spiez), universities, hos-
pitals, and the two Swiss Federal Polytechnical Schools ETH (ZUrich) and EPFL
(Lausanne). The radioactive substances may involve radio-nuclides for medical, indus-
trial, or research application. The radioactive sources may be sealed or unsealed, and the
radio-nuclides are classified according to their toxicity. Handling and storage of
radioactive sources is submitted to the Swiss regulation concerning the peaceful use of
atomic energy and the oIdinance on radiation protection (Figure 2.22).
• Atomic Law: "l.oi fedirale du 23 dicembre 1959 sur l'utilisation pacifique de l'energie
atomique et la protection contre les radiations"
• Radioprotection ordinance: "Ordonnance Concernant la Protection Contre les Radia-
tions, 30th of June, 1976"; ("Verordnung aber den Strahlenschutz vom 30. Juni 1976")
• SDRIRSD ordinance: "SDR, Ordonnance Suisse Relative au Transport des Marchandises
Dangereuses par Route, 1972)". The German terminology is: "RSD, Verordnung flir die
Beforderung geftihrlicher GUter auf der Strasse, vom 17April 1985".
Figure 2.22. Swiss regulation concerning the peaceful use of atomic energy and the ordinance on
radiation protection.
Transport authorizations are not required for radioactive sources, as far as their
activity is below 20 pica-curie per gram for alpha emitters, and 20 nano-curie per gram
for beta emitters. Special transport authorizations are required for transporting larger
quantities of radioactive sources. Any transport must comply with special packaging
requirements prescribed in the ordinance.
The authorities in charge of controlling the Swiss nuclear activitiel', the HSK,
"Hauptabteilung fUr die Sicherheit der Kernkraftwerke" is responsible of the application
of the ordinance and regulation and is c;lelivering transport permits. The international
transport of nuclear fuel pins from a nuclear power station to an intermediate storage or
to the nuclear reprocessing facilities of La Hague (France), as well as from a fuel
manufacture to a Swiss nuclear power station is taking place in special heavy transport
casks under police escort. Such transports are, of cause, subjected to special authoriz-
ations and must be announced to the Swiss Government (Department of Inland Affairs)
as well as to International Atomic Energy Agency (lAEA) in Vienna.
Transport Inside a Nuclear Facility. The ordinance prescribes also that for radioactive
sources transported inside an authorized facility but outside a controlled radioactive
zone, the dose should not exceed O.lmSvlh (10 millirem per hour) at a distance of one
meter from the surface and 2 mSvlh (200 millirem per hour) on the surface.
Transport Outside a Nuclear Facility. Transports of radioactive substances outside a
facility, i.e., on private or public roads, must comply with the RSD/SDR ordinance. The
maximum allowable activity for free parcels is limited accoIding to tabulated values,
whereas a distinction is made regaIding solids, liquids and radioactive gases. The
contamination on parcels, trucks, containers, tanks and other packaging material (which
is of non-adherent in nature) has to be as low as possible and is not allowed to exceed
the following values (Figure 2.23 and Figure 2.24).
ASSESSMENT OF TRANSPOTATION RISKS IN A COMPLEX AREA 81
Figure 2.23. Transports of radioactive substances outside a facility: - Parcels subjected to free traffic.
B) Parcels not subjected to free traffic with UN-Nr. 2912, 2913, 2918, 2982
Maximum allowable Dose per hour for parcel belonging to this class but not subjected to
supplementary restrictions:
• 2 mSvlh (200 mremlh) measured at the surface of the parcel
• 0.1 mSvlh (1Omremlh) at a distance of one meter from the parcel surface
Maximum allowable dose per hour for parcel belonging to this class but subjected to
supplementary restrictions:
• higher than 2 mSvlh (200 mremlh) but in any case less than max. 10 mSvlh (1,000
mremlh).
Figure 2.24. Transports of radioactive substances outside a facility: - Parcels not subjected to free traffic.
Maximum allowable sUrface contamination for Beta, Gamma and Alpha Emitters of low
toxicity (non-adherent radioactive confllmination):
• 0.4 Bq/cm2 (J O-S IlCi/cm2) for parcel subjected to free traffic and/or non radioactive
goods
• 4 Bq/cm2 (10-4 IlCi/cm2) for all shipment
• all other alpha emitters
• 0.04 Bq/cm2 (10-6 IlCi/cm2) for parcels subjected to free traffic and/or non radioactive
goods
• 0.4 Bq/cm2 (1 O-SIlCi/cm2) for all other shipment
Figure 2.25. Transports of radioactive substances outside a facility: - maximum allowable surface conta-
mination.
2.7. BUWAL's Guidelines for Assessing the Risks ofTraftic Ways in Switzerland
The Swiss Federal Office for Environment, Forestry and Landscape, BUW AL(i) has
issued a "Handbook No. Ill, to the Incident Regulations: - Directives for Traffic Ways
(December 1992)". This handbook complements the regulations and ordinance on
incidents and is to be used for assessing the risks of traffic ways2 (see below). The
methodological approach proposed by BUWAL for risk assessment, and the factors to
be used for risk calculation differ slightly from those methods described in our book.
We are, therefore, presenting in this Chapter, in a shorted form, the general and legal
aspects of Handbook No. III, whereas the BUWAL methodology for estimating the
specific risks of transportation of dangerous goods by road, rail, and waterways (i.e., the
Rhine river) will be presented in the respective chapters concerning such subject The
main topics from the original BUW AL handbook were translated from German into
English and reproduced in a summarized form to complement the chapters two to four
of this book.
The official BUWAL document in German, French, or Italian language may be
purchased from the Swiss Federal Office for Printed Matters and Office Material
(EDMZ(ii».
Based on the act from February 27, 1991 decreed by the Swiss Federal Government,
which came in effect on April I, 1991, the Swiss regulations concerning the protection
against incidents (so-called "Regulation Concerning Incidents", abbreviated in German
as StFV, 1991)42 came into effect on April 1st, 1991, and is based on Article 10,
paragraph 4 of the "Environmental Protection Law" (USGt 3 from October 7, 1983;
and Article 26 paragraph 1 of the "Waters Protection Law" (GSchG. 1991).44
Both regulations allow the Swiss Federal Government to issue prescriptions con-
cerning the safety measures of processing plants, by applying their right in the form of
a prescription.
Based on Article 39, Paragraph 1 of the "Environmental Protection Law (USG,
1983)", and Article 47, Paragraph 1 of the "Waters Protection Law (GSchG. 1991)",
the "Regulation Concerning Incidents (StFV, 1991)" describes more precisely different
tasks and duties mentioned under the Articles 10, Paragraph 1-3 and 44-47 of the USG
law, namely official surveys; control and information right; non-disclosure and informa-
tion duties.
(.) BUW AL. BWldesamt fiir Umwelt, Wald WId Landschaft. Bern. Schweiz.
(n) EDMZ. Eidg. Drucksachen- und Materialzentrale. EDMZ, 3000 Bern, Schweiz.
ASSESSMENT OF TRANSPOTATION RISKS IN A COMPLEX AREA 83
The objective of the Swiss "Regulation Concerning Incidents (StFV. 1991)" is the
protection of the population and the environment against severe damages, which might
occur as a result of (abnormal) plant operation or transportation incidents.
The main topics of this regulation are summarized below:
• Identification of risk concerning the population and the environment
• Risk reduction
• How the Owner has to manage the risk
• Control of the self-responsibility of the Owner by the Competent Authority
• Improving the information of the population concerned, in order for it to become
more aware of the risks involved.
The competent authorities are consequently allowed to issue prescriptions in the
form of handbooks to cover the following three domains:
1) Processing plants dealing with substances. products or special wastes
2) Processing plants dealing with micro-organisms
3) Traffic ways on which dangerous goods are transported
In the following chapters/sections we shall only be concerned with item (3), dealing
with the risk assessment of traffic ways on which dangerous goods are transported.
The Handbook III (Dec. 1992) to the "Regulations Concerning Incidents, StFV -
Guiding Principles for Traffic Ways" concerns more precisely the Owners of traffic
ways (railroads, transit roads and highways, and the Rhine river) where dangerous
goods are transported or loadedlunloaded.
''Transports Through Public Services" or the corresponding goods and objects defined
in international agreements. Each dangerous goods transport for which use is made of a
railway installation is to be considered as a transport and is consequently submitted to
the prescription.
The competencies and duties between the Swiss Confederation, the Cantons, and the
Executive Authorities were defined as Indicateri below:
Information and Alarm in Case of Incidents. It is the duty of the cantons to inform in
good time the population concerned and eventually to release the alarm and indicate the
appropriate measures to be taken. They have also to inform and eventually alarm the
neighboring cantons and states when the incidents might have severe consequences
across the cantons or states borders.
The accuracy of absolute risk results depends on whether all the significant risk
contributors have been included, the realism of the models, and the uncertainty
associated with the input data. If the principle risk contributors can be estimated from
adequate historical data, the resulting uncertainties will be small, but they can be one to
two orders of magnitude larger in case of rare events for which the statistical data are
scarce. This statement was confirmed by the result of a benchmark exercise prepared
during an a recent international conference on the risk of transporting dangerous goods
(Saccomanno, 1993).48
The substances chosen by BUW AL to be included in the Swiss study were judged to be
representative of those categories of substances which could cause large numbers of
fatalities or cause a severe environmental pollution, and which represented a significant
proportion of the overall traffic in dangerous substances.
Summary
References (Chapter 2)
22 Streblow, R.A., Unconfmed vapor cloud explosions - An overview, Presented at the 14th Symposium
(International) on Combustion at Pensylvania State University, (August 1972).
23 Westbrook, G.W., The bulk distribution of toxic substances: a safety assessment of the carriage of liquid
chlorine, in Loss Prevention and Safety Promotion, Vol. I, (1974), pp.197.
24 Browning, R.L., Estimating loss probabilities, Chern. Engng., Albany, Dec. IS, (1969), pp. 135.
2S Leimkiihler, F.F; Karsen, M.J.; Thompson, J.T., Statistical analysis of the frequency and severity of
accidents to potential highway carriers of highly radioactive materials, Report No. NYO-9771, U.S.
Atomic Energy Commission, (1961).
26 Leimkiihler, F.F, Trucking of radioactive materials: Safety vs. economy in highway transport, (ed.) The
John Hopkins Press, Baltimore, Maryland, (1963).
27 Chira-Chavaia, T., Data from TRB-proposed national monitoring systems and procedures for analysis of
truck accident rates, in Transportation Research Record, , Transportation Research Board, Washington,
DC, 1322 (1991), pp. 44-49.
28 Hu, P.S., Estmating commercial truck VMT of interstate motor carlers: - Data evaluation, Report ORNU
TM-11278, Oak Ridge national Laboratory, Tennessee, (1989).
ASSESSMENT OF TRANSPOTATION RISKS IN A COMPLEX AREA 89
29 Harwood, D.W.; Russel, E.R., Present Practices of Highway Transportation of Hazardous Materials, Report
FAWH-RD-89-013, U.S. Department of Transportation, (May 1990).
30 Abkowitz, M.; List, G.F., Hazardous Materialss Transportation: Commodity Flow and Incident/Accident
Information Systems, U.S. Congress Office of Technology Assessment, (January 1986).
31 Bendixen, L.M.; Barkan, C.P.L., Development of an interindustry risk model for hazardous material
transportation by rail. Paper read at the International Consensus Conference: Risks of Transporting
Dangerous Goods, April 6-8 (1992), University of Waterloo, Toronto, Canada (1992).
32 Purdy, G., Risk l:Ilaiysis of the transportation of dangerous goods by road and rail, J. Hazard. Mater. 33
(1993), pp. 229-259.
33 Hubert P.; et aI., Les risques d'accident majeurs dans les transports de mati~res dangereuse, Centre d'Etude
sur l'Evaluation de laProtection Nuc16aire, D 114, France, (1985).
:w HSC: Major hazard aspects of the tranport of dangerous substances: - Report and appendices, Healt and
Safety Commission (HSC), Advisory Committee on Dangerous Substances, HMSO, London, (1991).
3~ European Agreement concerning the International Carriage of Dangerous Goods by Road (ADR), and
protocol of signature, Economic Commission for Europe (ECElfRANS/I00) Inland Transport Commit-
tee, United Nations, New York, (1993).
36 Dangerous Goods Regulations, lATA Resolution 618, Attachment "A", International Air Transport Asso-
ciation, Montreal-Geneva, 1993.
37 Red Book on Transportation of Hazardous Materials, Second Edition, (eds.) Lawrence W. Bierlein, Van
Nostrand Reinhold Company, New York, (1988).
38 Cheremisinoff, N.P., Transportation of Hazardous Materials - A Guide to Compliance, (ed.) Noyes Publi-
cations, New Jersey, USA, (1984).
39 Kieselbach, R., Transport geflihrlicher GUter in der Schweiz, - Basisinformationen I Standortbestimmung
1992, Swiss Federal Laboratories for Materials Testing and Research, Diibendorf I ZUrich, EMPA
Bericht Nr.227, (1993).
40 Dangerous Goods Regulations, lATA, International Air Transport Association, 34th Edition, (lst January
1993).
<I Berger, H.U., Transport radioaktiverstoffe auf offentlichen Strassen, unter Beri1cksichtigung der dem KfK
erteiltcn Genehmigingen, Bericht KfK 5085, (Oktober 1992).
<2 Verordnung yom 27. Februar 1991 fiber den Schutz vor StOrflillen (SWrfailverordnung, StFV), SR 814.ot2.
43 Bundesgesetz Yom 7. Oktober 1983 fiber den Umweltschutz (Umweltschutzgesetz, USt;), SR 814.01.
.. Bundesgesetz VOD' 24. Ianuar 1991 fiber den Schutz der Gewlisser gegen Verunreinigung (Gewlisserschutz-
gesetz, GSchG), SR 814.20.
<~ Verordnung yom 5. November 1986 fiber den Transport im offentlichen Verkehr (Tranportverordnung,
TV), Anlage 1 (Ordnung fur die Schweizerische EisenbahnbefOrderung geflihrlicher Gfiter, RSD), SR
742.401.
46 Durchgangsstrassenverordnung yom 18. Dezember 1991, Anhiingen 1 und 2 .
., Verordnung Yom 29. April 1970 fiber die BefOrderung geflihrlicher Gfiter auf dem Rhein (ADNR), SR
747.224.141.
<II Saccomanno, F.F.; Yu M.; Shortreed, J.H., Risk uncertainty in the transport of hazardous materials, in
Transportation Research Record 1383, Transportation Research Board, National Research Council.
Washington, DC, (1993), pp. 58-66.
CHAPTER 3
General information and know-how related to truck accident rates, accidents in tunnels
and their consequences are presented in this chapter. The Swiss methodology recom-
mended by BUWALi for assessing the risks of hazaIdous materials transportation by
road is explained in details. Finally, new developments in traffic and vehicle control are
presented.
In the late 1960s and early 1970s, Brobse (USAEC 3 , 1972) developed a scheme for
categorizing highway and rail accident severity in terms of impact speed and fire
duration. TABLE 3.2 shows Brobst's values for accident frequencies associated with the
u.s. accident severity subcategories (USAEC 1972), which were based on Leimkuhler's
speed distribution.
TABLE 3.2. Accident frequency based on severity category
Since the early 1980s, it has been known that the truck accident rate varies with road
type and with population density. The most widely cited (TABLE 3.1) values valid for
the USA are those of Harwood and RusseI4 (1990). Both used U.S. data bases that
included data on highway geometries, truck volumes, and truck accidents. The above
mentioned weighted average values (TABLE 3.1) are proposed by FHWA as national
default values, that is, for use in U.S. risk analysis when better data are unavailable.
There is however a substantial variation in accident rate between some states an the
national U.S. average value (Hobeika and Kims, 1991). These authors pointed out that
the largest variations occur where the sample sizes were the smallest, and the weighted
average minimizes the effect of such values. Therefore the average values have an
uncertainty of as much as ±2. There are numerous earlier studies that substantiate the
RISK ANALYSIS OF HAZARDOUS MATERIALS TRANSPORTS BY ROAD 93
values in TABLE 3.1 (Smith and Wilmot6 , 1982; Abkowitz et al.7, 1984; Graf and
Archuleta8, 1985; Jovanis et al.9, 1989).
A set of data that are consistent with preceding (TABLE 3.1) are presented in
TABLE 3.3.
TABLE 3.3. u.S. Truck accident rates by highway type
TABLE 3.4. Partial listing offactors considered to affect truck accidents (USA)
A word of caution must be given here: not all database analyses are equally valid
The user of reported data analyses must develop its own confidence level with the
methodology used to analyze the data. One way of doing this is to compare the results
reported by several risk analysts. ldea1ly. the values reported in the literature should
describe any differences from previously reported values and explain the variations.
Sometimes the total number of accidents reported by some source may differ from the
94 CHAPI'ER3
Impact Speed of Heavy Goods Vehicles at the MOlMn' of a Collision (Great Britain)
• Built-up areas: A total of68 accidents, of which 36 were heavy goods vehicles collisions.
• Outside built-up areas: A total of 32 accidents, of which 21 were heavy goods vehicles
collisions.
• Motor ways: A total of 13 accidents, of which 7 were heavy goods vehicles collisions.
• The speed was ranging from 0 to 60 mph (96 km/h). It must be noted that in Great
Britain, the road conditions and habits of driving are different from those in Switzerland.
Figure 3.1. Impact speed of heavy goods vehicles at the moment of a collision (Great Britain).
under the Transportation of Dangerous Goods Regulations (SOR/85-77). Since that date
the Transport Dangerous Goods' Directorate of Transport Canada has collected data on
numerous characteristics of each reported spill in a comprehensive data base. IS
The 1986 and 1987 data for gasoline and liquefied petroleum gas (LPG) truck spills
that occurred en route were examined with the objective of calibrating the accident and
release submodels for use in a transportation risk model that was developed previously
for Transport Canada.
There are only a few statistical data and studies concerning the risk of accidental
releases and fIres in road tunnels, that were reported. In this section and in Chapter 10
we will review the results of some risk studies dealing with the risk of hazardous
materials transports in road tunnels.
Fire and Explosion Risk. The fIre and explosion risk of a hazardous material road
tanker in a highway tunnel is a function of the frequency with which an incident may
occur and the magnitude of such an incident. Fires resulting from the ignition of
hazardous substances may take on a number of forms including pool fIres and vapor
fires w.here the volatile liquid spillages (such as gasoline) are involved and, addition-
ally, for pressurized liquefied gas releases, torches (and fireballs). The frequency of
tunnel fire is usually very low, but the magnitude of such a fire in a highway tunnel is
significant. The pool fire mode will be by far the most important in terms of event
frequency. A pool fire inside a tunnel will cause heating of the walls by radiation and
convection from flames and hot gases. Significant damage may occur by thermal
spalling of tile grout and concrete. Usually, all fittings engulfed in the fire will be
damaged. Both thermal effects from flames and hot gases, and toxic effects of smoke
and combustion products will represent a threat to people in a tunnel fire. A fire
involving a 113.56 liters (30-gal) spill or a 75.71 Vmin. (20 gpm) leak of a liquefied
flammable gas or Class I flammable liquid, or involving a 605.7 I (l60-gal) spill or 100
gpm leak of a Class II or Class II combustible liquid, will endanger all people within the
tunnel but will probably not cause structural damage. On the other hand, a fire
involving a 378.54 liters (lOO-gal) or 151.42 Vmin.(40 gpm) leak of a liquefied
flammable gas or Type I flammable liquid, or a fIre involving 1.892.71 liters (500-gal)
spill or 757.08 Vmin. (200-gpm) i,~ of Class II or Class II combustible liquid, will
present a severe fire exposure to the tunnel structure, with ceiling temperatures ap-
proaching 1,093 °C (2,000 OF) for longer than 1 hour.
When a signifIcant vapor cloud builds up before ignition, a vapor fire is possible.
The hazards from vapor fires, in which significant overpressures are not involved, will
be dominated by the damage produced by flame contact of usually short duration. In
such a case cars present in the tunnel may take fire, but the tunnel structures will not be
damaged to a large extent. A hazardous material cargo spill involving a liquefied
flammable gas, or Type I flammable liquid that does not involve an immediate fIre can
create a significant explosion potential in a tunnel. The result of such an explosion
(subsonic deflagration or supersonic detonation) can create a blast overpressure that will
cause structural damage to the tunnel. Similarly, an explosion potential exists if a fire
involving a liquefied flammable gas or Class I flammable liquid is extinguished before
all of the available fuel is consumed or contained. Class II or Class III combustible
liquids do not present a significant explosion potential unless they are heated above
their flash point by an exposing fIre.
Quasi-instantaneous and continuous release of pressurized liquefIed flammable
gases will, when ignited almost immediately, bum as fireballs and torches, respectively.
RISK ANALYSIS OF HAZARDOUS MATERIALS TRANSPORTS BY ROAD 99
m o
·c 405
Societal Risks
(Frequency of N, or more fatalities per 10~ years)
Event
Motorists Public
Roule N=1 N=lO N=IOO N=1 N=IO N=IOO
Section
Corrosive/toxic liquid T 5.8 - - - - -
release OR 170 - - - - -
Flammable liquids T 45 15 15 - - -
OR 1,200 380 - 7 - -
Liquefied gas release T 1.6 0.61 0.44 - - -
OR 47 5.3 - 0.29 0.11 -
Toxic gas release T 3.2 3.2 0.19 - - -
OR 57 21 0.66 6.5 1.8 -
Asphyxiant gas release T 0.19 - - - - -
OR 6.6 - - - - -
Condensed phase OR 0.04 0.04 0.02 - - -
explosion T 0.88 0.44 - 0.12 0.8 -
Total 1,600 420 16 14 1.9 0
~: T: tunnel; OR: open road
A Swiss study, entitled MORIKA, concerning the risk presented by the "Baregg" road
tunnel (National highway Nl, Baden, Switzerland) has been prepared recently for the
Canton of Argovia. 22 The knowledge and results of this survey will be presented in
Chapter 10.
Based on article 5 paragraph 2a-d of the Ordinance of April 1, 1991 concerning the
"Protection Against Incidents" (Storfallverordnung, StFV)23 the Swiss Cantons have to
102 CHAPTER 3
Highways (1994)
Total 71,106 km; paved 71,106 km (including 1,502 km of expressways).
In Switzerland, there are still several highways and new roads in construction. As
soon as a longer segment is ready, it is open to the traffic. This explains the differences
found in above mentioned figures, anyway this give an idea of the situation.
In a first step, the cantonal road network is to be represented on a general map. For a
clear presentation of the statements asked for in the short report, the road track should
be subdivided into several road segments. The length of a segment is determined in
such a way, that the architectural and technical configuration, the environment, the
traffic and the safety measures should be homogeneous within the selected segment.
The length of a segment should not be less than a kilometer.
The data to be recorded are: population density, groundwater resources, surface waters,
as well as other objects worth considering. Furthermore one has to determine an
observation corridor, which shall be, in case of disturbances, subdivided into a close-
range compartment seriously affected, and a far-range compartment much less affected.
The close-range compartment extends 200 m on the left and on the right of the road
sides; the adjacent far-range compartment extends both sides up to 1500 m. The
indicated distances are standard values, which suffice as a rule for the recording of the
en'.'ironmental data as part of the short report. In case of a particular disposition of the
row in the terrain, or due to a particular environment (narrow valleys; groundwater
resources) it may be necessary to adjust these distances according to the circumstances.
For each road segment, the statements about the environment are to be tabulated, and
whenever possible, entered in a general map at a scale of 1: 25'000.
Population Density. The population density is to be indicated for the close- and far-
range along the road, and should mention the number of inhabitants per square
kilometer (Inh.1km2). If there is no data available, than the population density could
possibly be estimated as follows:
urban population density: > 5000 Inh.1km2
small-town population density: 2000 to 5000 Inh.1km2
village population density: 100 to 2000 Inh.1km2
slight or no settlement: < 100 Inh.1km2.
Groundwater Zone. On the basis of the available maps mentioning waters protected
against pollution, all protected groundwater zones are to be recorded in the close-range,
S. For each protected zone the unsaturated zone as well as the flow direction of the
groundwater should be estimated and, if possible, the importance of the groundwater
resource. Likewise, for each protected groundwater zone, one has to indicate the
104 CHAPTER 3
Surface Waters. In the close-range all surface waters (flowing waters and standing
waters) and their distance to the road are to be recorded, or if along the road, they are
part of the main drainage channel of the road segment under consideration.
In the far-range, the surface waters are to be recorded, if they can be polluted
through waters, which lie in the proximity of the road, or are acting as receiving waters
of a main draining segment under consideration. For the flowing waters the yearly
average discharge quantity (m3/s) is to be estimated; whereas for the standing waters it
is the water volume (m3). These data can be taken from the "Hydrological Year-book of
Switzerland" .
The waters are to be described shortly with regard to:
- the direct use for drinking water supply (i.e., direct drawing out the waters)
- the indirect use for drinking water supply (i.e., natural or artificial feeding into a
groundwater stream, from which drinking water will be taken)
the importance as fishing waters.
Objects to be Considered. The objects listed below will only be considered in the
close-range:
• Schools, homes for old-age people, market centers, sport yards, camping yards,
and exhibition halls.
• Cataloged objects that are of national interest, like protected environment/-
landscape, hill moor, pa~ture-Iand, tale moor, and moor landscapes.
• Cataloged large fruit cultures.
• Industrial processing plants with a potential danger.
• Water supply plants.
• Other important traffic ways.
RISK ANALYSIS OF HAZARDOUS MATERIALS TRANSPORTS BY ROAD 105
The traffic data described in the following section must be determined for each road
segment.
Traffic Rise
The traffic volume is defmed here, as the Average Daily Traffic per 24 hours (ADT-24).
The Heavy Traffic Share (HTS) corresponding to this ADT-24 must be indicated.
Average Daily Traffic per 24 Hours. This is defined as the yearly total of vehicles at a
certain road cross section divided through 365. This can be determined for certain
through-roads from the published statistics of the Swiss traffic counting, which takes
place every five years. Since such counts records catch hold of the ADT-24 data, a
conversion is necessary. If records concerning the average (hourly) traffic of motor
vehicles at day (Nd) and night (Nn ) are available from the cantonal noise-pollution
registers, then the ADT-24 data can be determined from such records.
Heavy Traffic Share of ADT-24 (HTS). The defmition of the heavy traffic includes all
trucks, tractor-trailer units, articulated lorry, and special-purpose vehicles (> 3.5 t total
weight, without touring bus). The Heavy Traffic Share, HTS, can be estimated like
ADT, because of the above- mentioned Swiss traffic counts; or alternatively, the Nd2 and
Nnl data about average traffic at day and night (or, hourly traffic) can be used instead.
These data are available from the cantonal noise pollution-registers concerning motor
vehicle traffic. When corresponding data are lacking, one can apply the Swiss mean
value of 6% to '!8timate the HTS value. This ratio is to be adjusted accordingly by
means of estimates to reflect the local/regional conditions prevailing on a given road
segment.
Traffic Composition
In order to assess the composition of the Swiss traffic, one has to indicate the Share of
Dangerous goods traffic based on the Heavy traffic (SDH) and the ratio of the different
SDR-classesi corresponding to the dangerous goods traffic (RSC).
Share of Dangerous Goods based on the Heavy Traffic (SDH): According to the latest
road traffic survey of the year 1984, the share of the dangerous goods traffic related to
the total heavy traffic amounts to 8% (Swiss average). Depending on road segment and
regional particularities this proportion can vary between 5% and 15%.
, SDR-class: All defined in the Swiss Ordinance on the transportation of dangerous goods on roads. (SDR,
Regelung tiber die schweizerische Beforderung gefabrlicher Gtiter auf der Strasse). See also ADR,
International Ordinance on the Transportation of Dangerous Goods on Roads. (ADR, Regelung tiber die
internationale Beforderung gefabrlicher GUter auf der Strasse).
106 CHAPfER3
Ratio o/the Different SDR-Classes based on the Dangerous Goods Traffic (RSC): For
transit roads, we can assume the following distribution, which is based on the Swiss
average (TABLE 3.10):
TABlE 3.10. Ratio of the different SDR-classes to the dangerous goods traffic (RSC)
SDR-class 1 2 3 4 5 6 7 8 9
RSC 0.001 0.07 0.70 0.07 0.Q1 0.07 - 0.08 -
Depending on the regional aspects (presence of a harbor, stores, loading/unloading
station, chemical plant, or processing plant) the RSC-share corresponding to a given
road segment may be corrected on the basis of estimates.
Accident Statistics
The Accident Rate of the Total traffic is to be considered (ARn. This is to be calculated
according to the Swiss VSS-directive24 (VSS-Norm SN 641'310a). When some accident
statistics are not available, the accident rates corresponding to different road categories
can be selected from the data listed in TABLE 3.11, showing the accident rates, and in
brackets, the confidence limits.
TABlE 3.11. Accident rates for the total traffic (Average values AR-total)
If statistics concerning Accident Rates for the Heavy traffic (ARB) are available for
the different road segments, they must be used for the calculation. Where this data are
not available, then the accident rate for the total traffic is to be multiplied by a factor
0.5. This is because the share of the heavy traffic accidents to the total traffic accidents
(Swiss average) is approximately hiM as large as the share of the heavy traffic to the
total traffic. In special cases, e.g., strong ramp, this value can be higher and an
appropriate correction must be made.
The methodology followed by the Swiss Authorities is based on the most actual
national and international knowledge and experience in this field. Due to some
simplifications the method cannot cope equally well with all kinds of situations, like for
instance very long tunnels. The method allows to coarsely assess, for each road
segment, the probability of an incident causing severe damage to people, groundwater
resources, and surface waters, on the basis of representative incident scenarios.
Severe Damages
For this study, the following figures defining severe damages have been adopted:
Population: 10 or more deaths.
Groundwater: the exceeding of the tolerance value of the food ordinance for mineral
oil products, or for greatly water-polluting fluids at a drinking water catchment
in a protected groundwater zone S.
Surface waters:
Running waters: the contamination(a) of a river over a length of 10 km, or
more.
Stagnant waters: the contamination(b) of a lake over a surface of 1 km2 , or
more.
The RSS-factors have to be adapted to local/regional conditions.
(a) Here the contamination is defined as the overstepping of a solids loading of a river (including bank zone)
of 15 glm2 through mineral oil products, and/or 5 g/m2 of greatly water-polluting fluids.
(b) Accordingly, it is assumed, that rivers of less than 10 km length, and lakes of less than I km2 area can not
be badly damaged. However, it must be considered that, by the intermediate of these smaller surface waters,
other rivers or lakes can possibly be polluted. In a such situation, one is to decide case wise, whether a
frequency should be determined for the larger river or lake, depending on their size. The RSS-values are to
be eventually adjusted depending on the local realities.
108 CHAPTER 3
RRl: It was assumed a rate of 0.04 for a relevant release (quantity> ca. 1.5 t), and a
probability value of 0.05 for an ignition following such a release. As a result RRl is
taking a value of 0.002.
• The RRP-value is based on data published in the annual report of the "Schweizerische ErdOlvereinigung",
(i.e., the Swiss Petroleum Products Association).
110 CHAPTER 3
RSS: The consequences of a fire extend rarely over the road breadth. As a result it is
the road users that will mostly be affected. The RSS-value (thermal radiation damages)
for road users, corresponding to high ADT-values on open road (0.3), were taken from
the HSC-study and adjuste4 to reflect Swiss conditions. The remaining values for RSS
were derived by this. RSS takes thus, depending on the average daily traffic (AD7) and
population density (PD), one of the values listed in TABLE 3.12.
For tunnel other RSS-factors are necessary, since incidents in tunnels lead to
somewhat larger consequences. Additionally, a distinction is to be made between
tunnels comprising one a tube, and tunnels with two tubes because the exposition to
heat radiation is different.
The values for RSS, valid in such a case, are shown in TABLE 3.13.
TABLE 3.13. Delennination of the RSS-faa.ors for the scenario "Fire"
(affecting the road users in a twmel)
RRI: A value of 0.002 was adopted for relevant releases (quantities> 2.0 tons) followed
=
by ignition. In this case RRI 0.002.
RISK ANALYSIS OF HAZARDOUS MATERIALS 1RANSPORTS BY ROAD 111
RSS: The effects of an explosion extend further than for a fire. Beside the road users, a
larger part of the population will also be affected. The value for explosion damage
concerning road users, corresponding to high ADT on open road (0.8), were taken from
the HSC-studrs and adjusted to reflect Swiss conditions. The remaining values for RSS
were derived from this value (0.8). RSS takes thus, depending on the average daily
traffic (ADn and population density (PP) one of the values of TABLE 3.14.
For "tunnels", the same procedure as for the scenario "fire" was adopted. The
corresponding values can be taken from TABLE 3.15.
RSS: The effects of a toxic gas release are concerning the road users, as well as the
residential population. When unfavorable conditions are encountered, one is to include
the far-range, as well, into the risk aspect considerations.
TABLE 3.16. Determ.ination of the RSS-factor for the scenario "Release of Toxic
Gas" (affecting road users and the population)
TABLE 3.17. Determination of the RSS-factor for the scenario "Release of Toxic
Gas" (affecting road users in a tunnel)
The RSS-values for toxic gas release concerning road users, corresponding to high
ADT on open road (0.6), were taken from the HSC-studis. The remaining RSS-values
were derived from this study and modified to reflect Swiss conditions. As a
consequence, RSS takes one of the values listed in TABLE 3.16, depending on the
average daily traffic and population density.
Especially with long tunnels, and in particular in urban regions, there is a problem
with aeration exhaust chimneys. This must be considered when computing a modified
value for RSS. The procedure adopted for tunnels was similar to that for fire. The RSS-
values are shown in TABLE 3.17.
RISK ANALYSIS OF HAZARDOUS MATERIALS TRANSPORTS BY ROAD 113
RRI: A rate of 0.004 was assumed for relevant releases of fuel oil (quantity> 10 tons).
Consequently RRI is equal to 0'()04.
RSS: Protected groundwater zones S are usually delineated in a way such that incidents
involving the release of mineral oil products (e.g., fuel oil) outside the protected zone
cannot conduct to a severe contamination of the groundwater close to the catchment
point. Therefore, the probability of an incident leading to severe contamination of the
groundwater resources is only to be estimated for protected groundwater zones S, which
are crossed by the road, or are affected by it
The RSS-factor can be found in TABLE 3.18. Fm the case that particularly efficient
measures have been taken for the prevention of water pollution, then RSS is to modified
accordingly.
TABLE 3.18. Determination of the RSS-factor for the scenario "Release of MineraI Oil
Products" (affecting groundwater)
Thickness So i I Permeability
of Soil Cover small medium high
<2m RSS = 0.05 RSS=0.2 RSS=0.5
2m-10m RSS = 0.01 RSS=0.05 RSS = 0.2
>lOm RSS=O.Q1 RSS =0.Q1 RSS=O.05
TABLE 3.19. Determination of the RSS-facror for the scenario "Release of Liquids Potentially
Very Noxious to Warer"
RRl: A mte of 0.02 was assumed for the release of relevant quantities of tetrachlorethen
(quantity> ca. 2 t). In this case RRI is equal to 0.02.
RSS: The RSS-factors are listed in TABLE 3.19. For the case that the road is built as a
an impervious sump, a coefficient for RSS of 0.01 should be used.
TABLE 3.20. Determination of the RSS-factor for the scenario "release of mineral oil products"
(affecting surface waters)
RRI (jor flowing waters): A rate of 0.02 was assumed for a relevant release of
CHLORETHEN (quantity> ca. 2 t). As a result RRI(Flowing Watenr) has a value of 0.02.
RRI (for lakes): Quantities smaller than 5 tons are not going to cause severe damage to
lakes. A rate of 0.005 was assumed for a relevant release (quantity> ca. 5 t). As a
result RRl(lAke.) has a value of 0.005.
RSS: The value of the RSS-factor for flowing waters is shown in TABLE 3.21, and for
lakes in TABLE 3.22.
116 CHAPTER 3
Traffic control systems, such as the railway signaling system and the traffic light control
system are basically complex distributed control systems.. These large scale systems
and networks are made up from a hierarchically structured subsystems which will
interact with one another. In order to control such highly complex systems, conven-
tional mechanisms have become inefficient and impractical. The reliance on sophisti-
cated computer control systems to manage such systems and networks is becoming
increasingly essential in order to match with ever growing demands and needs. New
developments in traffic safety management, software and computer system for vehicle
and traffic control are emerging.
hoped that a viable solution to current and foreseeable problems in European road
transport can be found. A major assumption that has influenced the direction of the
DRIVE project is the idea that whereas we would appear to be faced with an unlimited
demand to use the road networks by both private vehicles and increasingly by inter-
national freight, there is a limit, due to both political and environmental pressures, to
the number of new roads that can be built in Europe. A solution to this problem is seen
in the implementation of a new and effective information systems, navigation systems
and traffic management systems.
Within these three areas the following ideas are being seriously considered:
• Use of artificial intelligence.
• Automatic or semi-automatic vehicle control systems (so-called intelligent co-
pilot).
• Lane and positioning control and collision avoidance systems.
• Road usage and access control.
In view of the problematic of alternative routing in some parts of the country, the
implementing of vehicle and traffic control systems shall definitely help improving
traffic safety and reduce the number of accidents involving trucks or road-tankers.
Summary
This chapter introduces the methodological framework for performing a risk analysis of
hazardous material transportation by road in the context of regional risk assessment and
safety management. A large variety of .info~tion and knowledge about truck accidents
and associated consequences are needed for complex transportation risk analysis.
Special attention is given to the risk assessment of the transportation of hazanlous
materials through road tunnels: examples include data and cases from UK, USA, as well
as from Switzerland The Swiss methodology for assessing the risks of hazanlous
materials transportation by road includes the need for the subdivision of the road tracks
into road segments, various information and data concerning the environment such as
traffic rise, traffic structure and accident happening. It allows the estimation of the
likelihood of an accident with severe consequences to people and environment.
Scenario generation for risk assessment is of a fundamental importance; various
examples on how to design risk scenarios are included in this chapter (e.g., scenario for
surface waters, for release of mineral oil products, groundwater contamination, etc.).
118 CHAPTER 3
References (Chapter 3)
1 Leimkuhler, F.F.; Karsen, J.M.; Thompson, J.T., Statistical Analysis of the Frequency of the Severity of
Accidents to Potential Highway Carriers of Highly Radioactive Materials, Report No. NY0-9771, U.S.
Atomic Energy Commission, (1961).
2 Brobst, W.A., Transportation accidents: how probable? Nuclear News 16(5), (1973), pp. 48-54.
3 USAEC (U.S. Atomic Energy Commission). Environmental Smvey of Transportation of Radioactive
Materials to and from Nuclear Power Plants, WASH-1238, (December 1972).
4 Harwood, D.W.; Russel, E.R., Present Practices of Highway Transportation of Hazardous Materials. Report
FAWH-RD-89-013, U.S. Department ofTransportation, (May 1990).
S Hobeika, A.G.; Kim, S., Databases and needs for risk assessment of hazardous material shipments by
trucks, in Proceedings of Hazmat Transport '91 - A National Conference on the Transportation of
Hazardous Materials and wastes, Northwestern University, Evanston, Illinois, (1991).
6 Smith, R.N.; Wilmot, E.L., Truck Accident and Fatality Rates Calculated from California Highway
Accident Statistics for 1980 and 1981, Report SAND82-7066, Sandia National Laboratories, (1982).
7 Abkowitz, M.; Eiger, A.; Srinivasan, S., Assessing the Release and Costs Associated with Truck Transport
of Hazardous wastes, Report PB84-224468, U.S. Environmental Protection Agency, (1984).
8 Graf, VD.; Archuleta, K., Truck Accidents by Classification, Report FHWAlCAffE-85, U.S. Department
of Transportation, (1985).
9 Jovanis, P.P.; Chang, H.L.; Zabaneh, I., A comparison of accident rates for two truck configurations, in
Transportation Research Record 1249, Transportation Research Board, National Research Council,
Washington, OC, (1989), pp. 18-29.
10 Harwood, D.W.; Russel, E.R., Present practice of highway transportation of hazardous materials, FHWA-
RD-89-Ot 3, U.S. Department of Transportation, (1990).
11 Davies, P.A.; Lees, F.P., Impact speed of heavy goods vehicles, J. Hazard. Mater. 26 (1991), pp. 213-217.
12 Krasner, L.M., Motor Vehicle Standards for Hazardous Material Transportation, National Highway Safety
Bureau, Washington, OC, (Jan. 1970).
13 Jones, G.P., et aI., Risk Analysis of Hazardous Material Transportation, Vol. I, Report TES-20-73-4-I, U.S.
Department of Transportation, (1973).
14 Rhoads, R.E., et aI., An Assessment of tl>e Risks of Transporting Gasoline by Truck, Report PNL-2133,
Battelle Pacific Northwest Laboratory, Richland, Wash., (Nov. 1978).
IS Stewart, A.M.; Van Aerde, M., An empmcal analysis of Canadian gasoline and LPG truck releases, J.
Hazard. Mater. 25 (1990), pp. 205-217.
16 Considine, M., Risk assessment of the transportation of hazardous substances through road tunnels, in
Recent Advances in Hazardous Materials Transportation Research, An International Exchange, Confer-
ence of Lake Buena Vista, Florida, November 10-13, 1985. State of the Art Report 3, Transportation
Research Board, National Research Council, Washington, OC, (1986).
17 Transport Statistics Great Britain 1971-1981, Department of Transport, Her Majesty's Stationary Office,
London, England, (April 15, 1977).
18 Hills, P.J., The Relative Dangers of Carrying Conventional Fuels in Britain by Inland Service Modes, Oyez
International Business Communications Seminar on the Risk of Fuel Transport, London, England, (June
16-17, 1982).
19 Monitoring the Movement of Hazardous Freight by Road, Research Report 45, Transport Operations
Research Group, University of Newcastle upon Tyne, England, (1982).
20 Maclean, AD., 1980 Chemical Incidents Survey: Statistical Analysis, Home Office Scientific Advisory
1A VSS: Analyse der StrassenverkehrsunfaIle, Schweizer Norm SN 64I'31Oa (1990), Vereinigung Schwei-
zerischer Strassenfachleute (VSS).
2S Health & Safety Commission (HSC): Major hazard aspects of the transport of dangerous substances, Report
and appendices, HMSO, London, (1991).
26 Inter-Agency Program on the Assessment and Management of Health and Environmental Risks from
Energy and Other Complex Industrial Systems: Manual for the classification and prioritization of risks
from major accidents in process and related industries, International Atomic Energy Agency, Vienna,
(199\).
27 U.S. Federal Emergency Management Agency, U.S. Dep. of Transportation, U.S. Environmental Protection
Agency: Handbook of chemical hazard analysis procedures, Washington IX, (1989).
28 Fermaud, C.H.; Bohnenblust, H.; Biihlmann, B., Gewasserschutz und Transport gefahrlicher Gilter, Risiko-
analyse fiir die Nationalstrassen des Kantons Uri, SIA 46191, (1991).
29 Williams, M, "PROMETHEUS is Rolling", Colloquium on the Car and its Environment - What DRIVE
and PROMETHEUS have to offer, lEE, (January 1990).
CHAPI'ER4
In this chapter we are first reporting the findings on train accidents experienced in the
USA and Great Britain. Then we present information concerning the situation in
Switzerland, and the Swiss Federal Rail Networ~ (SBB/CFFIFFS) is briefly described.
The methodology proposed by BUWALii for assessing the risks of hazardous materials
transportation by rail is explained in details, and finally, selected results of a risk
assessment study prepared by SBB, that concern incident frequencies and the indicators
used to represent the extent of expected potential damages are presented at the end of
this chapter.
Train accident rates are more straightforward to obtain than those for trucks, because
there is usually one source of accident data, i.e., the railway companies collecting their
own accident data pertaining to a given railroad-network. In the USA the accident
reports must be submitted to the Federal Railroad Administration (FRA) by the
different railway companies and are then recorded in the RAIRS database (Railroad
Accident/incident Reporting System). Accident reports must be filled in when (a) there
is any impact with rail equipment and road users at a rail-road, respectively rail-
highway crossing, (b) the opemtion of the railroad results in death or injury, and (c) any
accident results in damages exceeding the reportable threshold.
The Association of American Railroad (AAR) also maintains a database (TRAIN IT)
on railcar movements. Actually 90% of the railcars are included, and 100% of their
movements and the commodity being carried are recorded.
A good source of information is the Accidentllncident Bulletin published annually
by the FRA Office of Safety (USA). TABLE 4.1, showing train accidents by track class,
has been taken from this FRA Bulletin. This organization publishes also an annual Rail-
Highway Crossing Accidentllncident and Inventory Bulletin.
For the calendar year 1991, the number of train miles amounted to a total of 5,768 x
108 , of which 0.88 x 108 for yard switching. An overall accident mte of 4.6 x lO-6/train-
mile (excluding gmde crossing accidents) can be computed for the same period, and an
i SBB: Schweizerische Bundesbahnen; CW: Chemins de fer fedCraux suisses; fFS: Ferrovie federali
svizzere (i.e., the Swiss Federal Railways).
h BUWAL: Bundesamt fUr Umwelt, Wald und Landschaft, Bern, Switzerland, i.e., Swiss Federal Office for
Environment, Forestry and Landscape.
122 CHAP1ER4
accident rate of 14.4 x 1O-6/train-mile is obtained for yard switching. The accident rate
without the yard or crossing values resulted in an overall rate of2.4 x 10-61 train-mile
for 1991.
TABLE 4.1. U.S. train accidents by traclc class, Year 1991
The AAR and the RPI (Railway Progress Institute), and the Chemical Manufacturers
Association (CMA) are jointly developing a hazardous materials transportation risk
analysis model. The accident cause module will include the effects of accident type
(collision, or noncollision), train speed. track class, railroad type, train length, ear
placement within the train, tank ear type, and car protective features (TABLE 4.1).
Further, the risk scenario has to consider the type of car involved (empty, loaded),
the availability of combustibles (type, amount), the probabilistic involvement of cargo
and of adjacent cars. From these data the bum rates can be calculated, and the resulting
fire duration estimated, taking also into consideration the fire-fighting efforts.
•••
Accident on rail-highway grade crossing
Derailment without collision
Collision
• with ctU'S remaining 011 trtu:k
• with del'tlilmeflt of ctU'S following the collision
0 over bridge, with car falling
- into water
- on clay, silt
- on hard soil, soft rock, or concrete
- on hard rock
- rail bed, roadbed
0 over embankment, with car falling
- into drain ditch
- on clay, silt
- on hard soil, soft rock
- on hard rock
0 into slope
- on clay, silt
- on hard soil, soft rock
- on hard rock
0 into structure
- column
- abutment
- other structures
0 rollover
... rollover followed by collision
- with locomotive
- with car
- with coupler
... rollover without collision
- on roadbed
- on earth
• Other causes
Figure 4.1. Train accident scenario used for estimating impact and crush fon:es.
124 CHAPfER4
The number of tank car failures from puncture can be calculated as follows:
RISK ASSESSMENT OF HAZARDOUS MATERIALS TRANSPORTS BY RAIL 125
(4.1)
where
Np =
Number of tank car failures from puncture, on a yearly basis
Mr =
Rail car mileage per year (mile/yr, or km/yr)
Fa = Accident frequency per km, or mile
Ppe =
Probability of a puncture environment
PPI =
Probability of puncture failure
The probability of large container puncture during rail accident has been estimated
by Denis et al. (1978) for U.S. railroads and is given in TABLE 4.2.
Clarke et al. 3 (1976) have analyzed the severity of transport accidents for a truck
trailer or railcar with a load of small radioactive material containers. It is worth
mentioning that the analysis should be also applicable to typical hazardous material
containers, e.g., 55-gal drums. The original data used by Clarke et al. (1976) and Denis
et al. (1978) can be validated by comparison with truck and rail data, as reported by
Fischer4 et al. (1987) and Harwood and Russel (1990). s
Certain dangerous substances, which are transported by rail in large quantities, have the
potential to cause situations presenting a major hazard of fife, explosion, or toxic
release: in particularly the possibility of killing large number of people in one event
For a risk assessment of transports by rail, a full consideration of all relevant dangerous
substances would require assessing a wide range of some hundred substances, which is
not feasible at a reasonable C,"lst The first priority is to concentrate on those substa.lces
which seem to present the greatest potential hazard, either because of the large bulk
transported (e.g., automotive gasoline, LPG), or because of the inherent hazard of the
substance (e.g., chlorine, ammonia).
The most common train accident is a derailment, in which a rail car separates from
the preceding car and leaves the track. A number of cars behind the lead car follow that
lead car. Train analyses is more complicated than truck analyses because the probability
that the derailed train segment includes the hazardous material of interest must be
considered.
In the early 1970s, Brobst6 (USAEC7 , 1972) developed a scheme for categorizing
highway and rail accident severity in terms of impact speed and fire duration. Brobst's
values for accident frequencies associated with the US accident severity subcategories
(USAEC 1972), which were based on Leimkuhler's speed distribution, have been
compiled in TABLE 3.1 (see Chapter 3.1).
Motor Gasoline
In Great Britain, motor spirit is transported in tank wagons with typical capacity either
32 or 75 tons, with nominal minimum shell thickness of 6 mm. In the period 1980-
1986, HSC 8 (1991) mention in their report that in 80 cases the spillage or leakage was
due to maloperation or failure of the tank wagon equipment and involved release of
126 CHAPTER 4
only small volumes of motor spirit. The other 4 incidents involved substantial spillage
as a result of derailment (3 events) or collision (one event). In one case 64,000 liters of
motor gasoline was lost.
Liquefied Gases
HSC (1991) mention in their reportS that, as far as they are aware, no liquefied gas tank
wagon has ever been punctured on railways in Britain. In the period 1980-1986 there
has been no sustained leak of LPG and no sustained leaks of toxic gases (ammonia,
etc.) .
Liquid Chlorine. The analysis of the risk arising from the transport of liquefied gases,
in Great Britain, has been reported by Purdy et al. 9 These authors describe the models
produced to analyze the level of societal risks arising out of the transport by rail of
chlorine.
Figure 4.2. Risk assessment approach for rail transportation of liquid chlorine in Finland.
Another assessment has been performed on the toxication risk of the population due
to bulk transportation of liquid chlorine by rail in Finland. 10 For this study fourteen rail
RISK ASSESSMENT OF HAZARDOUS MATERIALS TRANSPORTS BY RAIL 127
accident types were selected and the leakage probability was estimated for each type
separately (Figure 4.3).
In Finland the annual bulk. transportation of chlorine was about 150,000 ton by rail
and 200 ton by road in 1971. Trains hauling loaded chlorine tankers run about 500,000
kIn annually. There is on average one chlorine rail car in every 40 goods trains or one of
100 train is carrying loaded chlorine tankers. There are two size of tankers: 19-20 t and
45-50 t load capacity. The tankers are in Finland equipped with safety walls and jump-
over restraints.
The following information was collected: statistics with indication of number of
tracks, running times, switching times and places, number of crossing trains, number of
level crossings, density of motor traffic at the vicinity of such level crossings, etc. The
classification retained for railroad accidents was as shown in Figure 4.3.
Fourteen typical cases potentially affecting the integrity of the pressure tank were
retained for this study. Chlorine leakage were divided into four categories according to
the size of the vent and the liquid flow rates were estimated at 20 °C (TABLE 4.3).
Co"clw;wllS
The conclusions from the study on bulk transportation of liquid chlorine by rail in Finland,
show that:
• Derailment on the rail track does not represent a big /uwlrd.
• The leakage probability was highest for station accident involving a rail tanker and a
locomotive.
• Of the damage mechanisms considered, (i.e., valve group damages), impacts of general
type were found to dominate the leakage probabilities. The puncture mode was considered
relatively improbable as the safety equipment was assumed to protect the tank ends
effectively at collision speeds up to 40 kmIh. Head puncture by buffers and couplers of
adjoining wagons is generally the most probable damage type for unshielded tankers.
• In the smallest leakage category, random valve failure playa dominant role. The leakage
probabilities calculatedfor the Finish study can be taken from the reference paper.
128 CHAPI'ER4
An analysis of the British rail accident data regarding dangerous goods movements has
shown that there were two main causes of spillage: puncture or damage to the tank
wagon due to collision or derailment; and failure or maloperation of the tank wagon
equipment Spillage of motor spirit has resulted from both causes, but neither cause has
resulted in a sustained leak of chlorine, ammonia or LPG in GB from rail tankers.
In the British study the frequency of puncture of motor spirit tank wagons was
estimated from historical data, but leaks via tank wagon equipment were ignored. For
liquefied gases, Great Britain has experienced incident-free records over the last 16
years, thus the puncture frequencies were estimated using historical experience and
expert engineering judgment, and the equipment failure frequencies estimated using
fault tree analysis.
The leaks of liquefied gases due to tank wagon equipment failure were modeled as
two-phase releases from the appropriate valve. In the absence of any actual incident
data, the following assumptions were made:
a) For motor spirit tank wagons: it was assumed that 10% of the punctures would
result in an instantaneous release of the entire contents and the remainder would
be continuous spills from a hole with an equivalent diameter of 100 mm.
b) For liquefied gas tank wagons: it was assumed that 90% of the puncture events
would result in a crack or hole with equivalent diameter of 50 mm, and the
remainder wl)uld proceed to catastrophic failure. The likelihood and timing of
ignition of flammable releases was estimated using expert judgment based on the
historical incident data for flammable liquids.
The estimation of the hazardous consequences of spillage of dangerous substances
due to rail incidents was performed according to a methodology developed by HSFJ
HSCi .
a) It was assumed that spillage of motor spirit would result in a pool, which if
ignited would cause the deaths of all those within the areas of the resultant pool
fire.
b) It was assumed that spillage or leakage of LPG, if ignited, could result in the
following types of events: torch flame, BLEVEn fireball, vapor cloud explosion
or flash fire. The likelihood of each was determined by event tree.
c) RISKAT tool (developed by HSE) was used for assessing the risk of the
dispersion of a toxic gas for different weather conditions and scenarios. The area
within which people would be exposed to a specific toxic load was estimated and
this was related to the probability that a certain proportion of the exposed
I HSE: Health & Safety Executive, London; HSC: Health & Safety Commission, London.
iI BLEVE: Boiling Liquid Expanding Vapor Explosion.
RISK ASSESSMENT OF HAZARDOUS MATERIALS lRANSPORTS BY RAIL 129
population would be killed. Account was taken of the reduction in risk for
people being or going indoors.
The societal risk was estimated from the frequency of a particular event and the
number of people likely to be within the area where fatal consequences may be ex-
pected. The variation in trackside popUlation density along the route was considered,
and the risk to passengers in other trains along these lines was also estimated, this for
each group of dangerous products and main transport route. Finally, the national
societal risk from each group of substances was then estimated for both off-rail
residents and passengers, using these population densities, passengers movements, and
the total yearly movement of each substances. The aggregation of off-rail and passen-
ger risk is a complex process, and cannot always be produced by simple addition.
Upper Bound Puncture Frequency 1.4 X 10-7 per tank wagon kIn
Lower Bound Puncture Frequency 2.2 x 10-8 per tank wagon kIn
Mean Puncture Frequency Value 6.. 2 x 10-8 per tank wagon kIn
Ammonia and LPG Puncture Frequency 2_5 X 10-9 per tank wagon kIn
Chlorine Puncture Frequency 0.9 x 10-9 per tank wagon kIn
Using 1985 movements data the predicted frequencies of puncture and tank wagon
equipment failures per year, for Great Britain, were estimated to take the values of
TABLE 4.9.
TABLE 4.9. Estimated puncture and tank wagon equipment failures, valid for Great
Britain
Puncture Equipment
Failure Small
Leaks
For large spills of LPG where there is no immediate ignition there still remains a
chance that the flammable gas will encounter an ignition source and ignite before
dispersing down to the lower flammability limit. It was assumed that 70% of all large
spills would ignite giving an ignition probability factor of 0.70. Given that the assumed
132 CHAPTER 4
probability of immediate ignition is 0.2, then the probability of delayed ignition is 0.5.
In the case of small spills of LPG occurring on the railway we may expect the spill to
disperse rapidly, thus without causing delayed ignition.
TABLE 4.11. Number of fatalities after a pool fire event, as a function of pool area and
population density
TABLE 4.12. Predicted national risks due to rail transport of dangerous goods (for GB)
Rail Transport Estimated National Frequency per Year (x 10·')
for N greater than
Motor
Spirit
Off-rail 22,923 10,908 1,850 0 0 0 0
Passenger - - - - - - -
Total 22,923 IO,90t 1,850 0 0 0 0
LPG
Off-rail 1,020 687 522 235 82 24 0
Passenger 368 368 363 328 209 0 0
Total 1,299 1,059 923 653 365 24 0
Chlorine
Off-rail 362 197 174 87 82 20 13
Passenger 567 567 567 509 114 0 0
Total 600 591 587 576 243 66 14
Ammonia
Off-rail 1,596 1,090 910 896 304 33 0
Passenger 480 480 480 474 378 270 0
Total 1,892 1,398 1,288 1270 679 326 0
National (for four
Total substances )
Off-rail 24,902 12,882 3,456 1,218 468 76 13
Passenger 1,415 1,415 1,410 1,311 701 270 0
Total 26,714 13,596 4,648 2,499 1,287 416 14
(After HSC, 1991)
134 CHAPTER 4
The Swiss rail network is owned principally by the Swiss Federal Railway Company
(SBB, Schweizerische Bundesbahnen), but there are also some private railway com-
panies owning and/or operating more particularly railroads with special narrow size
tracks or cable trains.
The main rail network operated by SBB has the following characteristics (SBB I6,
1992):
• Total SBn rail network length: 2,973 km; of which
- 270 tunnels (211 km), and
- 4,114 bridges (58 km)
RISK ASSESSMENT OF HAZARDOUS MATERIALS TRANSPORTS BY RAIL 135
• Total installed rail length (i.e., length of all rail tracks of the network and
railroad stations): 7,391 kIn.
• Number of railroad stations, main and intermediate stations: 814 stations, of
which
16 are switching yards, and
- 28 are main switching freight yards centers (i.e., railroad stations in charge
of assembling freight trains)
• Number of road-crossing: 6,414, of which
- 3,261 are railway tunnels,
989 are railway bridges, and
- 2,164 are level crossings
The SBB owns a total of 2,156 locomotives, 4,496 railway carriages, and 20,236
freight cars. They rent 6,750 private freight cars. The maximum freight car load is 28
tons with two axles, and 60 tons with 4 axles. The dangerous goods traffic is principally
taking place on the whole rail network.
The SBB statistics 16 mention for the year 1992 the following:
• Passenger traffic
- Number of passengers transported: 260 million
Number of passenger' kIn: 11,034 million
Number of train· kIn: 28 million
• Goods traffic
Tonnage transported (tons): 51 million
Number of tons' kIn: 8,162 million
- Number of train· kIn: 28 million
• Dangerous goods according to RlDIRSD
Tonnage transported (20% of the total goods traffic
of 50 million tons): 10 million
of which:
- transported in freight trains: 6 million
- transported in single freight wagon: 4 million
The total number of trains involved in daily movements is shown below:
• Passengers trains: 4200 trains
• Goods trains: 2500 trains, of which:
- 25 trains transporting only petroleum products.
It must be noted, that a long-distance goods train includes, on average, a total of 20
wagons, each weighing 1,600 brut tons (1,100 net tons).
136 CHAPfER4
According to the same SBB statistics 16, the following rail traffic (in million net tons)
has taken place in 1990, totaling 10.1 million net tons:
- Inland traffic: 2.5
- Export traffic: 0.4
- Import traffic: 5.2
- Transit traffic: 2.0
The repartition of the goods was according to the following products groups:
- Petroleum products: 62 %
- Chemical raw materials: 25 %
- Gases: 4%
- Fertilizers: 3%
- Pharmaceuticals/dyes: 3%
- Others 3%
Based on rule 5 paragraph 2d of the Ordinance of April 1, 1991 concerning the "Protec-
tion Against Incidents" (StOrfallverordnung, StFV)17 the Swiss railroad company SBB
has to estimate the probability of incidents with severe injuries to the population or
severe damages to the environment.
Recently the Swiss Federal Office for Environment, Forestry and Landscape,
BUWALi has issued a "Handbook No. III, to the Regulations Concerning Incidents,
StFV - Guiding Principles for Traffic Ways" (December, 1992).17 This handbook is a
complement to the Ordinance of April 1, 1991:
We shall describe in this section the BUWAL procedure used for assessing the risk
of railroad transportation of dangerous goods in Switzerland, and discuss the main
topics to be covered by the "detailed risk assessment report" addressed to the competent
authorities (i.e., in this case BUWAL).
The general technical approach of the BUWAL methodology will be described below,
but some administrative and juridical details from the original report have however
been omitted here.
Environmental Aspects
The rail tracks considered must be subdivided into short segments (minimum 1 km in
length) with homogeneous properties, the length of which is determined by the
technical and constructional aspects, the surrounding environment, the traffic and the
safety measures related to a given segment. Among the constructional and technical
aspects of the rail network, consideration must be given to the location of a rail track,
type of track, and railway stations. Among the elements composing the track type we
find the track width and the number of tracks, railway platforms, bridges, viaducts,
galleries, and tunnels. Regarding the railway stations and switchillg yards, the length
between the entrance and exit of a switch, the surface covered and the approximate
movements of trains have to be mentioned. The safety aspects of the segment have to be
considered too.
The rail tankers, tank containers, and combined railIroad tankers ("Huckepack", i.e.,
transportation system involving the loading of road trucks/tankers on a wagon) have to
be assessed regarding number and capacity (in tons).
Along the rail track we draw a corridor subdivided into a close and a far range. The
close range has a width of 200 m on both sides of the track and the far range reaches
1,500 m on both sides. The data related to the surrounding environment along this
corridor must be compiled and comprise the following elements:
• groundwater resources and surface waters, as well as other object to be protected
• population density
Groundwater Resources: All the Swiss cantons have established g~graphic maps
showing the classification of the protected groundwater zones S (SI, Sll and SIll), as
well as zones of surface waters A, B, C, D. In the close-range, all protected groundwater
zones S must be mentioned. An estimation is to be made of the unsaturated zone
thickness (distance between soil surface and water table) and of the flow direction, and
quantity of ground waters. For the protected zones S one has to mention the planed
water poises and pumping capacity.
The permeability of the ground surface is to be roughly estimated, the value being
required for estimating the probability. The distance from the railway installation to the
water resources must be recorded as well as the usage (drinking water, water reserve,
fishing area).
138 CHAPTER 4
In the far-range, one is only to indicate protected groundwater zones S when they
might be indirectly contaminated through waters located at proximity of the railway
installations, or if the waters are acting as receiving waters of a drainage segment The
protected waters zones A may be indicated as well.
Surface Waters: In the close-range all surface waters (flowing waters and standing
waters) and their distance to the railway installations are to be recorded, if they are
located along the rail tracks, or if the waters are acting as receiving waters of the
drainage system of the railway installation (e.g., in tunnels).
In the far-range, the surface waters are to be recorded, if they can be polluted
through waters, which lie in the proximity of the railway installation, or are acting as
receiving waters of a main draining segment under consideration. For the flowing
waters the yearly average discharge quantity (m3/s) is to be estimated; whereas for the
standing waters it is the water volume (m3). These data can be taken from the
"Hydrological Year-Book of Switzerland". The intended use of these waters is to be
described: direct use as drinking water resources (i.e., direct drawing out), or indirect
use for drinking water production (i.e., natural or artificial feeding of a groundwater
table)
Special objects located in the close range, that need to be protected, are listed below.
These include for instance:
• Schools, hospitals, shopping centers, sport yards, camping, exhibition areas
• Protected objects or landscape zones of national relevance
• Inventoried fruit plantations
• Industrial plants with special danger potential
• Driuking water supplies
• Important traffic ways
Traffic Composition
The traffic composition regarding dangerous goods (net tons per year) is to be estimated
for each track segment. We distinguish between normal dangerous goods trains (miXed
with passenger cars) and special -dangerous goods trains (containing exclusively
dangerous goods), container traffic, and rail/road traffic combination ("Huckepack"
transport system).
Actually it is not possible to distribute the total quantity of dangerous goods
according to the RSD danger classes, so we have to use product categories instead.
Regarding the conventional goods traffic the Swiss average values can be taken from
TABLE 4.13.
The product categories 1231130 (Mineral Oil and Petroleum Products) comprise
only daf1gerous goods and represent in general the largest percentage quantity of
dangerous goods transported. When the quantities belonging to this product group are
known, then it is possible to distribute the rest of the dangerous goods transports
between the other product categories according to the ratio of the percentages for the
RISK ASSESSMENT OF HAZARDOUS MATERIALS TRANSPORTS BY RAn. l39
Swiss' average. The same consideration holds true for container and "Huckepack" traffic
(product categories 201 and 202).
TABLE 4.13. Distribution of dangerous goods according to product categories
Where possible actual quantities should be used and in case of necessity the
estimates should be corrected to reflect particular conditions (e.g., harbor, strategic
petroleum products storage, main transit axis or regional railway lines etc.).
For every dangerous goods transported by special train the average quantities (net
tons per train) and the number of special tft!.:lSports per year must be indicated.
Accidents History
Based on available incident protocols, the incidents/accidents must be evaluated for the
last ten years' period, and classified by accident type (i.e., derailment, collision, etc.)
and by accident location (i.e., open-air track segments, tunnels, railway station, and
shunting yards). Accidents involving dangerous goods, that resulted in leakage, fire, or
release of toxic gases or water-polluting liquids, must be mentioned with all details.
Basically, the rail accident rates must be calculated on the basis of the number of
derailments and collisions per traveled train·km, for open tracks, tunnels, and railway
stations. When data are missing, the following accident rates can be assumed to be
prevailing for Swiss conditions (year 1993): 0.12 x 1O-6/trains·kID for open tracks and
tunnels, and 0.6 x 10-6 Itrains·kID for railway stations. For trains using special narrow
tracks higher accident rates have to be assumed: i.e., 0.3 x 10-6 Itrains·kID for open
tracks and tunnels, and 1.5 x 1O-6/trains·kID for railway stations.
140 CHAPTER 4
The competent Swiss Authorities have fixed the following limits of acceptance
concerning the "incidents with severe consequences":
• Population: 10 or more people killed. ii
• Groundwater: any excess of the tolerance value, as defined in the Swiss Food.
Ordinance, either for mineral oil products, or for liquids that are potentially very
dangerous to water in the surrounding of a drinking water well located in a
protected zone S.
• Surface waters (rivers): the contaminationiii of a river over a length exceeding
10 km or more.
• Surface waters (lakes and ponds): the contamination of a lake iii over a surface of
1 km or more iv.
, Here the term "probability" is not defined, as usual, in the probabilistic sense. but has to be understood as
meaning "frequency".
U This defmition includes as well the population in the surrounding of railroad installations. as the passengers
RRI: A rate of 0.05 was assumed for a relevant release of gasoline (quantity> ca. 1.5
t), and a probability value of 0.33 for an ignition following such a release. As a result
RRI is taking a value of 0.1. For tunnels the value of RRI(tJUlMI) was assumed to be 0.5.
Reference Substance. FUEL OIL (for domestic heating) was chosen as representative
product. In order to estimate the value of YDG, one has to take 100% of the dangerous
goods of the product group "Mineral Oil Products". It will be assumed that the freight
transported per rail tanker is equal to an average of 55 t.
TABLE 4.17. Detennination of the RSS factor for the scenario "Release of
Petroleum Products"
Soil Permeability i
RRI for flowing waters: A rate of 0.25 was assumed for a relevant release of fuel oil
(quantity> ca. 19 t). As a result RRI has a value of 0.25.
RSS: The protected groundwater zones S have, in Switzerland, been chosen in such a
way that incident with petroleum products taking place outside the protected zone will
not lead to severe damages to the groundwater wells. As a consequence the probability
of an incident leading to severe damages to the groundwater resources will have to be
assessed only for protected groun.::1water zones S when the rail track is crossing the zont"
or is located very close.
The corresponding RSS-factors are listed in TABLE 4.17. In cases where particularly
efficient measures for groundwater protection have been taken, the RSS-factor is to be
adjusted in consequence.
, The soil permeability is given by the coefficient k. This coefficient is taking the values shown in TABLE
4.17.
146 CHAPfER4
TABLE 4.18. Detennination of the RSS factor for the scenario "Release of Strongly
Waler- Polluting Liquids"
Soil Permeability
Thickness of small medium high
Soil Cover k< 10.5 mls lO's < k < 10.3 k> 10.3
(fine sand, silt, (Clayey gravel (gravel)
clay stone, Sand )
Reference Substance. "FUEL OIL" (for domestic heating) was chosen as represent-
ative mineral oil product. In order to estimate the value of YDG, one has to take 100%
of the dangerous goods of the product group "Mineral oil and Petroleum Products". It
will be assumed that the freight transported per rail tanker is equal to an average of 55 t.
TABLE 4.19. Determination of the RSS-Factors for the scenario "Release of Petroleum
Products Into Lakes"
RSS=O.3 I RSS=O.l
RRI (jor flowing waters): A rate of 0.24 was assumed for a relevant release of fuel oil
(quantity> ca. 10 t). As a result RRI(Flowing Waters) has a value of 0.25.
RISK ASSESSMENT OF HAZARDOUS MATERIALS TRANSPORTS BY RAIL 147
RRI (for lakes): A rate of 0.15 was assumed for a relevant release (quantity> ca. 50 t). As
a result RRI(l-aus) has a value of 0.15.
RSS: The value of the RSS-factor for lakes is shown in TABLE 4.19, and for flowing
waters in TABLE 4.20.
TABLE 4.20. Detennination of the RSS-factors for the scenario "Release of Petro-
leum Products into Flowing Waters"
<50m 50m-200m
Without evident infiltration
of SBB I6 to BUWAL are briefly summarized below. The data are valid for the Swiss
rail network in general, but detailed risk assessment studies have also been prepared by
SBB along the same line, for each single rail track segments.
For the period 1979 to 1989 the SBB have registered around 460 incidents involving
dangerous goros movements. This figure includes aU incident registered by SBB, that is
also the leak of a fuel oil tank, as well as incident with SBB owned road tankers. In
70% of all the cases, the product involved was a petroleum product, and in 30% it was
another chemical product. 13% of the incidents resulted in a fire involving a petroleum
product. Most of these incidents did not endanger people or the environment, and the
resulting damages remained reasonable. In 3% of all the incidents, there were people
injured slightly, and in less than 2% a temporary pollution of the waters had to be taken
into consideration.
It must be noted that during the period 1990-1995 some severe rail accidents,
involving dangerous goods transport as well as passengers have taken place in
Switzerland (and also in other countries). This means that the probabilistic figures
prepared by SBB at the request of BUWAL have changed. Such probabilistic data
should any how be reconsidered every year as a function of the number of traffic
accidents/incidents.
Incidents with personal injuries (dead, evacuated persons) through the. transport of
dangerous goods are rare on the Swiss rail network. Thus their freqt.·ency cannot
reliably be determined only on the basis of the Swiss railroad statistics, because these
are in such a case insufficient due to the small sample size. Therefore, in the case of
SBB, the estimation of the incident frequency with personal injuries is prepared
principally on the basis of international statistics related to accidents involving
dangerous goods transportation by railroad.
Regarding incidents with environmental damages, international statistics do not
offer an appropriate support since corresponding data are hardly available. On the
contrary several incidents have happened in Switzerland in the past, where large
quantities of dangerous/toxic liquids have run out. These form a certain statistical basis
for the estimation of incident frequency with environmental damages. Indeed for most
incidents which led to a pollution of surface waters, groundwater, or soil, or had the
potential for a pollution, some quantitative statements about the extent of' the resulting
consequences are totally missing. Therefore the estimation of the extent of environ-
mental damages took place for the most part with the help of analytical model repre-
sentation.
The incident frequencies presented in the following pages represent values summed
over the entire SBB rail network, respectively the shares of the different track types.
They are valid for the present technical and operational factors of the SBB network
RISK ASSESSMENT OF HAZARDOUS MATERIALS TRANSPORTS BY RAIL 149
(reference year 1990). With regard to the calculation of incident frequencies for
local/regional rail tracks, the data must be put in relation to some characteristic values
like transport capability, traffic rise, population density, as well as the local situation
concerning surface waters and groundwater.
We show below (Figure 4.4), as example, the scales used by SBB for connecting the
different kinds of damages (Le., number of fatalities, people evacuated, contaminated
surface areas) with the event severity scale (defining acceptable and severe damages).
Figure 4.4 shows the definition adopted for the damage indicators used for assessing
the effects of an incident and the definition of classes measuring the extent of effects. A
numerical representation of the annual incident frequencies for the entire rail network is
given for all indicators and effects severity classes. These values are briefly interpreted
and commented below and in the next section.
The SBB have calculated that, on the Swiss rail network, one may expect a
dangerous goods accident once in every 40 years;, that will cause fatalities. Accidents
with severe damages, i.e., such one occasioning more than 10 fatalities, shall take place
once in 150 years. Incidents with more than 100 fatalities are essentially more seldom to
expect. For all three damage severity classes about 75% of the incidents are concerned
with fatalities occurring inside, and 25% outside the SBB railroad installations.
Once in every 8 years one has to reckon with an evacuation of people in the vicinity,
i.e., outside of the railroad installations. Once in every 40 years we may expect an
accident that will cause the evacuation of more than 1,000 people. Catastrophic
-\l~cidents that will cause the evacuation of more than 10,000 people should occur less
ClaD once in 300 years.
It is difficult to describe the concept of contamination exactly. Here, it was assumed
that we have to do with contamination, if at least short-term damages occur to fauna or
flora, or the utilization of the affected object by people is no more possible, at least
temporarily, or its use is very limited. The duration of the environmental damage,
however, is not considered explicitly for the assessment.
Contamination of groundwater catchment areas following an incidental leakage of a
dangerous good can be expected to take place once every 100 years. It was assumed that
in approximately 70% of all cases it was concerning a catchment with a pumping
capacity of less than 500 Umin, and in ca. 30% such one between 500 and 5,000 Umin.
Incidents causing severe damages to groundwater catchment delivering more than 5,000
Umin are expected to take place only once every thousand years. The latter value does
include the case where several catchment are contaminated at the same time.
IThis value corresponds to the average recurrence period. and is mathematically equal to the reciprocal
value of the corresponding frequency (e.g .• 1111 = 11 0.025 = 40 [years».
150 CHAPTER 4
Damage
Fatalities
1 10 100 Number of Fatalities
I I
Evacuations
1'000 10'000
I
Number of Persons Evacuated
I
I I
Surface Waters
0.1 1
I I
Groundwater Catchment
500 5'000
I I I
Figure 4.4. Example of a scale system used for assigning single damage classes indica-
tors to an event severity scale.
frequencies are an order of magnitude smaller and consequently not relevant. The
damage potential of accidents involving dangerous goods is, for the soil, comparatively
small. As opposed to the other indicators, the frequency of incidents that result in severe
soil damage are negligibly small, and thus can be omitted.
TABLE 4.21. Frequency of incidents classified by damage type and effects severity class for selected
indicators
accessible, and the narrow passages between railroad material and tunnel walls
(TABLE 4.22).
TABLE 4.22. Distribution of fatality incidents by track segment, inside SBB railroad installations
(F: Frequency per year)
Shunting yard 5% 2% -
Open track 20% 15% 10%
TABLE 4.23. Distribution of fatality incidents by track segment, outside SBB railroad installations
(F: Frequency per year)
Fatalities occurring outside SBB facilities (TABLE 4.23) are due to dangerous
goods accidents taking place mainly at railway stations. The reason is primarily due the
higher population density in the vicinity of a railway station, as compared to the
population density along the open railroad track. In the vicinity of shunting yards, the
probability of accidents involving many fatalities outside SBB railroad installations is
comparatively seldom. Indeed the velocity of the wagons during shunting is small com-
pared to usual railroad traffic speeds. Therefore the impact energy during collision or
derailment is generally to small in order to create severe accidents. Dangerous goods
accidents in tunnels are occasioning fatalities outside SBB facilities only in special
situations. Such accident type represent only a small share.
Most accidents necessitating evacuation of people can be expected to take place at
railway stations or on open railroad tracks. Tunnels and shunting yards contribute each
by about 10%, Other SBB statistics are given in TABLES 4.24 through TABLES 4.26.
RISK ASSESSMENT OF HAZARDOUS MATERIALS TRANSPORTS BY RAn.. 153
TABLE 4.24. Distribution of incidents by track segment. 00 SBB railroad insta1latioos. leading to
an evacuatioo of people (F: Frequency pel' year)
Tunnels 9% 4% 30%
· . the framework of the Incident Ordinance (1992)
Source: SBB short report within
TABLE 4.26. Distributioo of incidents by track segment, on SBB railroad installations. leading to
a contamination of surface waters (F: Frequency per year)
As already mentioned, the statistics and frequency data are no more up to date siBce
the publication of the SBB report. The risk assessment process, like a rolling-ff1reCast,
should be reviewed every year and revised accident frequency data prepared m order to
keep up to date with new developments concerning equipment modernization and safety
aspects on the rail network and in the tunnels. Railroad traffie frequency is also
changing over the years, and new rail tracks or switching yards are being implemented.
During the last years, there has been problems due to intensive rain-storm occasioning
earth movements that dragged away the tracks, or earth avalanche covering up the
tracks with mud, gravel and rocks. The risk of such event, which may have significant
consequences, like derailment, should be considered in future railroad risk assessment
studies. Recent earthquake in the USA and Japan have also damaged highways, rail
tracks and bridges of modem construction type. In Switzerland, such incident might
lead to land slide as well.
Summary
This chapter introduces the methodological framework for performing a risk analysis of
hazardous materials transportation by rail in the content of regional risk assessment and
safety management Experience with train accidents in a number of countries is
presented (e.g., USA, UK, Switzerland). Details are given on the Swiss methodology
for assessing the risks of hazardous materials transportation by rail: technical aspects of
the detailed risk assessment, estimation of the probability of an incident with severe
consequences for the population or the environment, various assumptions concerning
the representative incident scenarios involving the population (e.g. scenario for fire,
explosion, release of toxic gases, etc.), or the representative incident scenario for
groundwater resources, surface watt:rs, etc. The information and risk estimates pre-
sented in the last part of this chapter were compiled by SBB in accordance with the
requirements of the existing Swiss legislation on risk assessment for rail transportation
of dangerous goods.
RISK ASSESSMENT OF HAZARDOUS MATERIALS TRANSPORTS BY RAIL 155
References (Chapter 4)
1 FRA (Federal Railroad Administration), Accidentsllncidents Bulletin No. 160: Calendar Year 1991,
FedernI Railroad Administration, Office of Safety, Washington, D.C., (July 1992).
2 Denis, A.W.; Fowler, G.F., Railroad Tank Car Safety Assessment, Report RA-12-4-58, Association a
American Railroads, Chicago, (1990).
3 Clarke, R.K., et al., Severities of Transportation Accidents. Report SLA-74-0001, Sandia National Lab0-
ratories, (1976).
4 Fischer, L.E., et ai., Shipping Container Response to Severe Highway and Railway Accident Conditions,
Report NUREG/CR-4829, Lawrence Livermore National Labontory, (1987).
5 Harwood, D.W.; Russel, E.R., Present Practices of Highway Transportation of Hazardous Materials, Report
FHWA-RD-89~13, U.S. Department of Transportation, (1989).
6 Brobst, W.A., Transportation accidents: how probable 1, Nuclear News 16(5), (1973), pp. 48-54.
'USAEC (U.S. Atomic Energy Commission). Environmental Survey of Transportation of Radioactive
Materials to and from Nuclear Power Plants, Report W ASH-1238, (December 1972).
8 Health & Safety Commission (HSC): Major hazard aspects of the transport of dangerous substances, Report
and appendices, HMSO, London, (1991).
9 Purdy, G.; Campbell, H.S.; Grint, G.C.; Smith, L.M., Analysis of the risk arising from the transport a
liquefied gases in great Britain, J. Hazard. Mater. 20 (1988), pp. 335-355.
10 Lautkasi, R.; Mankamo, T., Chlorine transportation risk assessment, Technical Research Center of Finland,
Nuclear Engineering Laboratory, Report 27. Available from Valtion Painatuskeskus, PI.. 516, 00101
Helsinki 10, Finland, (1977).
11 Davies, P.A.; Lees, F.P., Accident speed of freight trains, J. Hazard. Mater. 28 (1991), pp. 367-370.
12 Prince, A. I., A users manual to SPIlL, UKAEA Reliability Directorate, Culcheth, SRD Report R2IO,
(1981).
13 Shaw, P.; Briscoe, F., Evaporation from spill of hazardous liquids on land and water, UKAEA Reliability
Directorate,Culcheth. SRD ReportRl00, (1978).
14 Mimer, G.A.; Eyre, I.A., Large-scale LNG and LPG pool fues, in The Assessment of Major Hazards,
Rugby, Institution of Chemical Engineers, (1982), pp. 147-164.
15 Opschoor, G., Theoretical study of the evaporation of liquids on land, Chapter 5 Ev~on, TNO Report
No. 78~8345, Nederlandse Organisatie voor Toegepast-Natuurwetenschappelijk ~derzoek Apeldoom,
The Netherlan<!s, (1978).
16 SBB: Schweizerische Bundesbahnen, Kurzbericht dec SBB im Rahmen der SWrfallverordnung, Teil I:
Netzweiteaspekte, SBB-Generaldirektion, Bern, (12. MiIrz, 1992).
17 BUWAL, Handbucb m zur SWrfallveroIdnung, Ricbtlinien fUr Verkehrswege, Dezember 1992, BWKlesamt
fUr Umwelt, Wald und Landschaft, Bern.
18 Health & Safety Commission (HSC): Major hazard aspects of the transport of dangerous substances, Report
and appendices. HMSO, London, (1991).
19 Inter-Agency Program on the Assessment and Management of Health and Environmental Risks from
Energy and Other Complex Industrial Systems: Manual for the classification and prioritization of risks
from major accidents in process and related industries, International Atomic Energy Agency, Vienna,
(1991).
20 US Federal Emergency Management Agency, US Dep. of Transportation, US Environmental Protection
Agency: Handbook of Chemical Hazard Analysis Procedures, Washington DC, (1989).
21 Verordnung tiber den Schutz vor SWrflillen (SWrfallverordnung, StFV, vom 27. Februar 1991).
CHAPTER 5
5.1. Risk Resulting from Ship and Barge Transport on Lakes and Waterways
Ship and barge transport on lakes and waterways present some risk to the environment,
people and to fixed installations and buildings. The greatest risk arises from explosion
of a flamrTUlble gas {propanelbutane, gasoline, solvents) cloud escaping from tanker-
ship damaged by a collision. Non-combustible gases, such as chlorine or ammonia, can
escape from a container as the result of mechanical or corrosion damage and form a
toxic gas cloud drifting in direction of populated areas along the lake or river. Another
risk may be due to water intrusion into a damaged barge and subsequent leaching of
some water soluble toxic substance, which may pollute lakes and rivers, thus presenting
a danger to ground-water and aquatic life.
operation on tanks or smoking near-by may result in fire and explosion. A cause of
explosion is quite often due to the presence of static electricity. The electrical charges
are created during the pumping of flammable liquids, as a result of a missing earthing or
the use of unsuitable piping equipment in conjunction with a to high a pumping
velocity.
5.2.1. OBJECTIVES
The objectives of an overall risk assessment of the major hazard aspects of the bulk
transport of dangerous substances by sea, in the ports and around the coasts, or on
inland waterways (Le., rivers and lakes) should be as follows:
1. To analyze the risks of major hazard accident arising in a given country as a
result of the transport of dangerous substances in bulk by ship or barge. The
risks to be assessed should include both individual and societal risks of death
due to fire, explosion, or toxic release. For shipments on inland waterways, the
analysis should include statements on the possible pollution of water surfaces by
such shipment, in case of fire or shipwreck.
2. To identify the main components of these risks.
3. To develop a methodology for such risk assessment adapted to the local
conditions, and suitable for evaluating remedial measures.
4. To evaluate selected remedial measures using cost-benefit analysis.
The main categories of dangerous materials which are transported in bulk include:
Crude oil
- Flammable and toxic liquefied gases
- Flammable liquid petroleum products
- Flammmable liquid chemicals
- Hazardous solids (e.g., ammonium nitrate)
- Solid and liquid chemicals potentially hazardous to water and aquatic life.
Concerning the transport of hazardous materials by ship or barge on inland
waterways, substances known to be potentially toxic to water and aquatic life should be
included in the study, since they can pollute a river, like the Rhine from Basle to
Rotterdam, as the toxic load moves down the river.
Many shipments are in tank containers, Ro-Ro vehicles, drums, and containers
enclosing smaller packages. Chemicals which are transported in small shipment on
parcel chemical carriers should be included in the study, provided that they are pumped
and not lifted ashore. Bagged ammonium nitrate, should be included because its
explosive effects are similar to those of the same material in bulk. The analysis of
TRANSPORTATION OF DANGEROUS GOODS BY SHIP 159
explosions of ammonium nitrite shows, however, that these are extremely unlikely
events.
A paper on risk assessment of shipping operations has been recently published by the
Health and Safety Commission (HSC) in Great Britain! and the methodology followed
is presented below. Three of the 88 terminals in 42 ports in Great Britain were analyzed
in detail and a simplified method was developed for the others. 2 "Major hazard
accidents" are accidents which threaten people on-shore or in other vessels in the port,
or crew members while the ship is berthed in port. Only spills or explosions of LPG are
likely to reach the population at risk. LPG is usually imported by both large fully-
refrigerated and small semi-refrigerated gas carriers. The small vessels have cylindrical
pressure tanks, but carry the cargo fully refrigerated. Their cargo capacity is typically
3,700 m 3 • The large gas carriers have prismatic tanks and a total capacity of 23,000 m3 •
The cargo is carried fully refrigerated at atmospheric pressure. An articulated arm with
quick-release and emergency shut-down is used for ship-to-shore transfer. Petroleum
products are imported in small tankers using flexible hoses for transfer.
The risk analysis of individual ports follows the classical fonn of quantified risk
analysis, consisting of the following steps:
• System definition, in which information on the ports and waterways and its
hazardous trades is collected.
• . Hazard identification, in which events are identified which may lead to the
release of hazardous cargo.
• Frequency estimatico, in which the frequency (i.e., likelihood per year) of each
event is estimated, u!tually from historical accident data.
• Consequence modeling, in which the consequences (i.e., fatalities) of each event
are estimated, usually by a mixture of accident experience and theoretical
models.
• Risk representation, in which the frequencies and consequences of each event are
combined to give the overall individual and societal risks.
• Assessment of the tolerability of the risks using agreed individual and societal
risk criteria.
The modeling of the consequences can be perfonned, for example, using the
computer program SAFETI (Ale and Whitehouse3, 1986), which allows efficient calcul-
ation of the effects of population distribution, meteorological conditions, ignition
source location, and release type, duration and location.
The risk presentation should follow the following fonn:
Individual Risk. The likelihood of death per year for an individual continuously at a
given location, assuming an average vulnerability to the hazards and allowing for
realistic probabilities of being sheltered indoors.
160 CHAPTER 5
F-N Curves. They are a measure of societal risk showing the relationship between
numbers of fatalities and frequency (i.e., likelihood per year) of accidents. The FN
curves are actually cumulative plots, showing the frequency F of accidents involving N
or more fatalities.
Annual Fatality rates. They are simpler measures of societal risk, equal to the expected
long term average number of fatalities per year (i.e., expectation values). These give no
indication of the numbers of fatalities in individual accidents.
Mean Fatalities per Event. They represent the average number of fatalities in given
accident types. These supply the missing information in the fatality rates. They are
equal to the fatality rate divided by the event frequency.
be spread anywhere over the entire width. According to this description effect models
should incorporate a correction factor or a different probability function.
The analytical process of estimating waterway vessel casualty probabilitiess
involves the following approach:
• Define waterway and subzones
• Analyze subzone historical vessel casualties
• Estimate subzone vessel traffic by type and size
• Quantify subzone waterway characteristics
• Calculate national average casualty rates by casualty type, vessel type, and size
(in the case of rivers, e.g., the Rhine, include traffic and accident data for the
whole river length, i.e., from Holland to the lake of Constance (Bodensee)
• Make regression analyses of subzones casualty rates with waterway characteris-
tics
• Develop subzone navigational risk adjustment values
• Finally calculate subzone specific vessel casualty probabilities by casualty type,
vessel type, and size.
The waterway specific characteristics may be described as waterway physical mea-
surements as well as meteorological and hydrologic conditions. Among the physical
measurements, we have channel width, depth, route alignment in terms of number of
turns, average or sum of total degrees of course changes, number of obstructions (e.g.,
bridges, anchorage, crossing lanes, etc.). total waterway length and total area of water
,surface in the subzones. These attribute should be measured for each river segment
Among the meteorological and hydrologic factors to be included are wind speed,
visibility, current velocity, water volume in m3/s. Having these information on hand it is
then possible to develop a linear multiple regression model for estimating the
probability of vessel casualties. The most significant variables to be selected for the
stepwise linear regression are: the waterway/channel characteristics, a local traffic
density factor, and two subzone type dummy variables (like constricted waterway,
harbor area, etc.).
It has been proven difficult to find a relationship between the meteorological or
hydrological conditions and the casualty probabilities, although visibility, wind speed,
and current velocity play a role regarding the navigational risk. In fact actual
meteorological conditions are more important than annual average values.
of flammable liquid around the jetty, tanker and ferry bow. Fire may spread from the
tanker to other vessel or ferry boats or harbor installations.
• Estimate the frequencies for the occurrence of the different scenarios, as explained
above.
• Assume different local meteorological conditions (a combination of wind speed
and atmospheric stability).
• Estimate the population density, using two distributions: one for day-time and one
for night-time. The data may be compiled from census data and from information on
the local industries in the region of interest.
• Establish the ignition sources distributions in order to model the probabilities of
ignition from single people, motor vehicles, industries, trains, ships, loading/filling
station, etc. based on judgment.
• Using special models like, for instance, SAFETI 3 (Software for the Assessment of
Flammable, Explosive and Toxic Impacts), calculate the spreading of the pool on
waterlland and the evaporation rate as well as the dispersion of the gas cloud with
respect to time. 6
• Calculate the possibility of ignition during the dispersion phase of the cloud, using
the lower flammability limit (LFL), and assuming a flash fire upon ignition in 90%
of the cases and an explosion in 10% of the time (vapor cloud explosion is very
unlikely). Different possible time steps for the occurrence of ignition must be
considered. Assume 100% of death within the reach of the flash fire for people
outdoors, and 30% for people indoors.
• Calculate the fatalities for each failure case and each of the atmospheric parameter
combinations, wind direction, population distribution (at day- and night-time),
ignition sources, time steps, and the proposed two ignited outcomes (flash fire and
vapor cloud explosion). The fatalities are expressed both as probabilities at each
point on a 100 m grid and as the total expected number of deaths.
• Compute the measures of individual and societal risk, using the probabilities of
each combination and the failure case frequencies. Usually the catastrophic
consequences are restricted to the vicinity of the harbor installations, but in some
instances catastrophic gas releases may travel over 1 krn before reaching their lower
flammable limit (with the wind prevailing in this direction), and so their frequency
is relatively low.
Based on the experience from different harbors, the largest contributions to the
societal risks arise from tanker transfer spills, which are ignited by passing motor crafts
or cars and from collisions between tankers and ferries or other ship. Sensitivity test can
be used to assess the effects of the proposed remedial actions. Finally using cost-benefit
analysis, it is possible to compare the costs of implementing remedial measures for
reducing accidents. The benefits of remedial measures are expressed in terms of the
risk-factored costs of accidents avoided.
TRANSPORTATION OF DANGEROUS GOODS BY SHIP 163
The frequencies which are required for a study are likelihood per year of an accidental
release or explosion of hazardous cargo on. Tanker explosion frequencies can be
estimated directly from historical data. The frequencies of cargo release should be
estimated for collision, striking, impacts and gas tanker fires as the product of two
164 CHAPTER 5
SF - YF (5.1)
- PM·PY
where
SF : Striking frequency (per passing movement)
PM : Passing movement (per year)
PY : Period of time (number of years)
Gas carrier Vessel struck Cargo tal Angle of Penetration Result of ProbabiHty
tanks
'struck at in way of loaded collision of tank accident
jetty cargo (obUque)
0.38 (Yes)
-
0.85 (No) No spill
0.5 (Yes)
-
Figure 5.1. Event tree for gas carrier striking.
Figure 5.1 shows, as an example, an event tree for a gas carrier striking, and the
associated event probabilities.
Expressing the frequency in this way allows it to be compared with striking
frequencies in other ports or fluvial segments with different traffic levels. The striking
frequency for a ship oflength equal to the jetty head can be estimated as:
_ SL TB·YF
~ - SF· JL' 24.365 (5.2)
where
F. : striking frequency for a particular ship (per year)
SF : striking frequency per passing movement
SL : ship length (m)
JL : jetty length (m)
TRANSPORTATION OF DANGEROUS GOODS BY SHIP 165
EDF (Electricite de France) has developed a method for estimating the risk resulting
from ship transport of dangerous cargo in the vicinity of industrial plants located along
a river or lake, based on probabilistic and deterministic approach. 7 The methodology is
described below.
166 CHAYfER5
The most general expression for the annual probability P j (year l ) of a fluvial or sea
accident, which can cause damage to people, property, or more specifically, to a
chemical or nuclear plant, is written, for a given product, i, and accident scenario s, as:
(5.3)
where
Pa : rate of occurrence of any accident on a lake or on a waterway per ship/barge
carrying dangerous materials and per kilometer of river bank or shore
p. : conditional probability of occurrence of an accident on a lake or on a fluvial
waterway, which could cause an explosion
p. : conditional probability of explosion
Pk : a weighting coefficient, associated with the geographical location studied and
the nature of the dangerous substance
F, : yearly traffic in the neighborhood of sites expressed in terms of the number
of ships carrying the dangerous substance, i, over a year
P, : probability of occurrence of the scenario, s
P!fW : probability of occurrence of given meteorological conditions on the site, as
defined by atmospheric stability and wind speed (sw)
Ej : factor associated with meteorological conditions (direction and frequency of
winds, conditions of atmospheric diffusion), which exists only in the case of
gaseous cloud dispersion
L, : length (km) of the river bank, or shore, or ship trajectory on a lake to be
taken into consideration:
(5.4)
Accident Rate
The accident rate per dangerous ship can be estimated from statistical data. As an
example we indicate typical data (Hendrickx et al. 7 , 1983) for French coastal waters
and sea passage (estimated for the year 1983).
Coastal waters: P a = 7.6 X 10-8 accident. (dangerous shipr!. (k:m of coa~t\ine)"!
In sea passage: P a = 2.9 X 10-6 accident. (dangerous ship)"!. (Ian of channel bank)"!.
These averages are weighted by the geographic location of the accident: PI allows
for this adjustment and was varying for the French study from 0.2 to 2.8.
The accident frequencies, in Great Britain, have been reported by the National Port
Council (NPC 8, 1976) and are reproduced in TABLE 5.2
TABLE 5.2. Ship accident frequencies in Great-Britain for the year 1976.
Wide estuary 4.0 x 10-5 8.0 X 10-6 4.0 X 10-6 2.2 X 10-3
Wide river 1.2 x 10-4 5.0 x 10-5 9.0 X 10-6 2.1 X 10-3
Narrow river 5.0 x 10-4 1.6 x 10-5 4.2 x 10-5 6.5 X 10-3
Similar accident frequencies for barge and ship traffic on the Rhine in the Basle area
must be calculated.
• The gas carrier must be struck in way of a cargo tank, since not all the length of
the ship is occupied by cargo tanks. Assume a probability of 0.65 (depending on
vessel type and tank location and spacing).
• The tank of the gas carrier must be loaded: assume a probability equal to 0.5.
• The collision must be at an oblique angle, with a probability estimated to 0.38,
and,
• The striking ship or barge must have sufficient energy to brea(;h the cargo tank.
This energy is proportional to the striking ship displacement and the square of its
speed, and is also affected by the strength of the other tanker structures, jetty, etc. In a
British study, the probability for a ship to have enough energy as been estimated to be
on average 0.31 for large gas carriers and 0.85 for small gas carriers. These values
should be reconsidered for the conditions prevailing on the Rhine in the region of Basle
(i.e., vessel type and size. speed, etc.).
The major hazards from bulk substances are shown in TABLE 5.4.
TABLE 5.4. Hazards from bulk transport
Material Outcome
Toxic liquefied gas Gas cloud
Rammable liquefied gas Vapor cloud explosion (VeE)
Rash fire
Fireball
Rammable liquid Explosion (of vapor)
Pool fIR'
Bulk solid Explosion
• the rate of flow from the breach and the volume discharge depend on the
position of the breach relative to the water line.
• the substance discharged forms a liquid layer on the water surface, and is being
vaporized due to heat supply from the environment.
• first a circular expansion, and then regression, of the layer, whose thickness
varies in space and over time up to a limit beyond which the effect of gravity is
taken as negligible.
• the vaporization caused by the supply of heat from ambient air, solar radiation,
and from the water (mainly for sea or lake water).
• the vaporization is further influenced by: convection phenomena, phase changes
(freezing of the water due to the evaporation process), conduction (the thickness
of ice varies with space and time).
t.I Refrigerated Catastrophic Contents release onto water over a period of 5 min.
Tanks cold rupture Predict spill vapor dispersion.
Cold leak from Assume continuous release of the liquefied gas is driven
pressurized tank by the internal pressure. Calculate th~ steady-state
or loading arm flashing release rate. The momentum of the jet is assumed
to be dissipated by the ship secondary containment
(double hulled walls for gas tankers). Predict spill vapor
dispersion.
172 CHAPTER 5
2.1 Refrigerated Hot rupture of a) To be modeled eitber as a fireball, on tbe basis that
Tanks tank containing pressure could build up inside the tank before it failed in
refrigerated a fire, or as a pool fire. Determine the thermal radiation
liquefied gas from the mass involved.
2.2 Pressurized Hot rupture of Assume a fireball and determine the thermal radiation
Tanks tank containing from the mass involved.
pressurized
flammable
liquefied gas
3.2 Full-bore Full-bore trans- The liquid flow will emerge at the pumping rate.
Leak fer spill during
loading or un-
loading opera-
tion
3.3 Pipe Leak Leak during For leak during transfer, assume an outflow at 10% of the
loading or un- pumping rate.
loading opera-
tion
It is assumed that 90% of cold releases are due to leaks and 10% to ruptures.
Experience has shown that using two different hole sizes for modeling would be
preferable, since it is now believed that a cold rupture of a gas carrier tank is very
unlikely. if a fire, however, causes rupture of a tank containing pressurized flammable
liquefied gas, this must be modeled as a fireball. The corresponding formula, which can
be found in literature, must be used. A model like SAFETI9 (which is basically an
improved version of the standard box model of Cox and Carpenter lO (1980) may be
1RANSPORTATION OF DANGEROUS GOODS BY SHIP 173
used to estimate the effects and consequences of (dense) gas release, flash fire, fireball,
or vapor cloud explosion. Possible scenarios assumptions are listed in TABLE 5.6
below.
Cold release of flammable liquid from a damaged tanker will be driven by
hydrostatic head. A constant flow rate equal to 75% of the initial rate may be assumed
in this case. This flow rate will continue until hydrostatic equilibrium is reached.
During transfer operation, the flow is assumed to emerge from a ruptured pipe at full
bore (loading arm) at the pumping rate. For a leak during transfer, assume a rate of 10%
of the pumping rate.
The release of liquid is considered to spread under gravity until it reaches an average
thickness of 5 mm. Then the spill will drift away under the influence of wind and water
currents.
Solid Materials Stored in Containers. Such containers are usually boxes, textile or
craft-paper bags, barrels, or steel drums. The storage method usually consists of piling
or storing these containers. Solid bulk materials are stored in piles, or in bins, silos or
bunkers located on the ship. Such solid materials present less problems.
TABLE 5.7. Scenario for water pollution by soluble (toxic) solid substances
Bulk Storage of Liquid Materials. This type of storage requires the use of reservoirs and
tanks. Liquid materials stored in containers are generally packed in barrels, kegs, drums,
cans, or glass or ceramic containers. These are beeing replaced, nowadays by plastic
containers. Cans are usually stacked on palettes.
TABLE 5.8. Scenario for water poUution by soluble (toxic) liquid substances
in which
p. : rate of occurrence of any accident on a lake or on a river per ship carrying
dangerous materials and per kilometer of river bank
P_ : conditional probability of occurrence of an accident on a lake or on a river,
which could cause a pollution of the waterway by a water soluble (toxic)
substance
P,oI : conditional probability of the material container being damaged, thus
releasing water-soluble substances (liquid or solid)
p. : a weighting coefficient associated with the geographical location studied and
the nature of the dangerous substance (liquid, solid solubility, toxicity class)
F, : yearly traffic on the waterway of concern, expressed in terms of the number
of ships carrying the dangerous substance, i, over a year
P, : probability of occurrence of the scenario s.
The scenarios resulting in water pollution of waterways as a result of transporting
dangerous goods by barge/ship are summarized in TABLE 5.7 and TABLE 5.8.
Accidental pollution of waterways during loading and unloading at a tanker terminal
or as a result of ship cleaning operations in the harbor have not been taken into
consideration in the scenarios mentioned below.
TRANSPORTATION OF DANGEROUS GOODS BY SHIP 175
Setting limits for acceptable risk, assessing existing risk levels and predicting the cost
and benefits of proposed safety measures are all meant to assist the risk management
process. If local factors have led to a relatively high accident frequency, improvements
to the infrastructure (waterways, bridges, locks) or the local traffic situation is the
logical way to look for a solution. But local improvement may not suffice: a reduction
in accident frequency of say 20% as a result of structural improvements or traffic
control is quite an achievement, but is sometimes not enough to reduce risks that
increase on a logarithmic scale. There are not too many alternative waterways which
would allow to choose another route and shifting to train or truck transport may not
improve safety. One has to look for a more generic solution at national or even
international level. It must be kept in mind that structural improvement to vessels will
normally only be feasible for newly built ships. From the current developments it can be
anticipated that in future all hazardous cargo will be transported in double hulled ships
(at least for ocean going ship). Furthennore the improvement of human and
organizational factors will playa significant role. It is important to define and establish
unifonn transport risk limits for all transport modes, i.e., not only for ship transporta-
tion. Once these limits are set it is possible to guarantee some generally accepted safety
level.
Navigation on the Rhine river all die 'year round has only become a practical possibility
in this century and necessitated regulation of the fairway on the upper Rhine. The first
towed barge arrived in Basle on June, 1904. Cargo-handling was makeshift but work
was soon started on the construction of proper dock facilities.
Constructional phases of the ports in the canton of Basle-Town:
SUohan (left bank) 1906-1911.
- Kleinhiiningen (right bank): Klybeck- and Rhine quays, Basins I 1919-1926;
Basin II: 1936-1942;
Constructional phases of the ports in the canton Basle-Country:
- Birsfelden and Au (left bank): 1937-1940 (extended after 1950).
The following information describes briefly the situation:
- Inland waterways
65 kID; Rhine (Basle to Rheinfelden, Schaftbausen to Bodensee);
12 navigable lakes
- Ports
Basle (river ports of both Basle)
176 CHAPTER 5
- Merchant marine
23 ships (1,000 GRT or over) totaling 337,455 GRT/592,213 DWT; bulk
10, cargo 4, chemical tanker 5, oil tanker 1, roll-onlroll-off cargo 2,
specialized tanker 1.
The ship traffic amounts to a total of 8,530,910 tons for the year 1992, respectively
8,353,401 in i991. The distribution of the goods was as indicated in TABLE 5.9 and
TABLE 5.10. The number of containers handled in 1991 amounted to 24,500 units
TEU (water side), respectively 60,000 units TEU (raiVroadiwater).
The transport capacity for Switzerland represents approximately 5 x 109 tons.kIn (or
ca. 60% of the goods transported by SBB-rail in 1991, which amounted to ca. 8.1 x 109
tons.km). This represents also 15% of the total Swiss foreign trade, which is handled by
ship transport on the Rhine.
The Swiss BUWAL (Federal Office for Environment, Forestry and Landscape) has
not requested a special risk assessment, investigation for the ports of both Basle, nor
have they proposed a particular methodology, but instead they request a short report
based on general risk considerations.
TRANSPORTATION OF DANGEROUS GOODS BY SHIP 177
TABLE 5.9. Goods traffic by product category, at the ports of both Basle
Coal 108,923
Crude oil , 20
Fertilizers 173,691
Summary
The chapter introduces the methodological framework for performing a risk analysis of
hazardous materials transportation by ship in the context of regional risk assessment
and safety management Techniques for risk assessment of ship are outlined within the
methodology for assessing the risk of individual ports. A computational framework for
the accident frequency estimation is included. The method developed by EdF identifies
178 CHAPTER 5
the annual probability of a fluvial or sea accident with possible damages to people, or to
a chemical, or nuclear plant. Various calculation procedures for the pollution of water-
ways by dangerous cargo, and the risk management for waterways and ship cargo are
considered in this chapter, too. Examples and information concerning the ports of both
Basle (Switzerland) are also included.
TRANSPORTATION OF DANGEROUS GOODS BY SHIP 179
References (Chapter 5)
1 Health and Safety Commission (HSC). Major Hazard Aspects of the Transport of Dangerous Substances,
Advisory Committee on Dangerous Substances, (1991).
2 Spouge, 1.R., The Use of risk assessment for ships carrying hazardous cargo in port areas, Conference on
Safety at Sea and in the Air - Taking Stocks Together, Royal Aeronautical Society, London, (November
1990).
3 Ale, B.I.M.; Whitehouse, R.J .• SAFETl: - Computer based system for risk analysis of process plants. in
Heavy Gas and Risk Assessment m,. Hartwig (ed.). D. Reidel, Dordrecht, The Netherlands. 5 (1986).
4 Hauer, Accident Anal. & Prev. 18 (1986), pp. 1-12.
S Spouge. I.R.. Techniques for risk assessment of ships carrying hazardous cargo in port areas. in F.
Saccomanno and K. Cassidy (eds.), Transportation of Dangerous Goods: Assessing the Risk, JnstilUte for
Risk Research (1RR). (1993). pp. 154-182.
6 Opschoor. G .• Methods for the Calculation of the Physical Effects of the Escape of Dangerous Materials.
TNO. The Netherlands Organisation of Applied Scientific Research, (1979).
7 Hendrich. S.; Lannoy, A .• Probabilistic and deterministic safety study on the sea-borne traffic of liquefied
gases in the vicinity of a nuclear site. in Loss Prevention and Safety Promotion in the Process industries.
IChemE 4th International Symposium on. 12-16 September. Harrogate. England. Vol. I - Safety in Opera-
tions and Processes, EFCE Publication Series No. 33 Vol. I. (1983). pp. DI-09.
8 National Ports Council (NPC. 1976). Analysis of Marine Incidents in Ports and Harbours. ISBN 0 90
1058939, (September 1976).
9 SAFETl (Software for the Assessment of Flammable. Explosive and Toxic Impacts): Ale. B.I.M.;
Whitehouse, R.J .• Computer based system for risk analysis of process plants. in S. Hartwig. D. Reidel
(eds.), Heavy Gas and Risk assessment m. (1986), Donlrecht. The Netherlands.
10 Cox. R.A.; CarPenter. R.J., Further development of a dense vapour cloud model for hazard analysis, in S.
Hartwig and D. Reidel (eds.). Heavy Gas and Risk Assessment. Dordrecht, The Netherlands, (1980).
CHAPTER 6
The natural gas consumption for the year 1992 was worldwide around 2'150 Mrd m 3
(1 Mrd = 1 billion = 109), which corresponds to a primary energy consumption of 23%.
In Western Europe the natural gas consumption for the same period reached 320 Mrd
m 3 (i.e., 14% of the world consumption). This corresponds to primary energy consump-
tion of 17%. The need were covered to about two third by supplies from Western
Europe own production and to one third by supplies from Russia, Algeria, and Libya.
The most important production site in Western Europe is Norway (North-Sea), with
2'000 Mrd m 3 of assessed reserves. The supply from this region amounts to 30 Mrd m 3
per year, of which one part is distributed by pipeline to Great-Britain and the other to
Continental Europe, corresponding to 9% of the consumption of Western Europe. The
second largest producer are the Netherlands Groningen), with assessed reserves of
1,950 Mrd m3 natural gas. The natural gas is exported from Holland by "Gasunie" and
amounts to 40 Mrd m 3 per year. The natural gas pipeline network in continental Europe
has a length of over 600'000 kID and shows many ramifications, allowing a fine distri-
bution to most important consumers (Figure 6.1).
Since 1977, the "Swiss Gas" company is importing natural gas from the North-Sea via
the German supplier "Ruhrgas" and since 1974 from the Netherlands through
"Gasunie". Actually Switzerland is inter-connected to ten different pipelines entering
the country in the North, East, South and West direction, thus ensuring a safe and
regular supply. Switzerland is also a producer of natural gas, since a gas field was
discovered in 1990 in Finsterwald (TABLE 6.1).
In Switzerland the natural gas network, including distribution network, total 12'143
kIn, of which 4'548 kID form a high pressure pipework and 7'595 kID a low pressure
distribution network.
The main Swiss natural gas transport pipelines totalize a length of 1'886 kID. This
figure includes the length of the transit pipeline from the Netherlands to Italy, which has
on the Swiss territory a length of 165 km. Part of this pipeline has been installed on the
ground of the lake of Lugano (Agno - Capolago), other parts (32 kID) are also installed
in tunnels in the Alps region. The length of the primary Swiss Gas network amounts to
244 kID and that of the regional distribution network 1'477 kID (Figure 6.2).
182 CHAPTER 6
Natural Gas Import and Consumption in Switzerland for the Year 1992
GWb TJ
The diameters and lengths of the main pipelines installed on the Swiss territory are
summarized in TABLE 6.2 (gas pipelines) and TABLE 6.3 (oil pipelines).
TABLE 6.2. Diameters of gas pipelines on the Swiss territory (after "Eidg. Rohrleitungsinspektorat",
Report SVDB No. 12411992)
Diameter 36" 34" 16" 14" 12 J14 " 10,J14" 8"·" 65/8 ' 4 112 • .3/8"
2
(Zoll)
Diameter 914.4 863.6 406.4 355.6 323.8 273.0 219.1 168.3 114.3 60.3
(mm)
Length (Ian) 93.8 70.7 254.3 48.9 228.4 484.4 252.7 113.2 64.3 0.8
The total installed pipelines lengths are: 1'611.5 kIn for gas pipelines, and 249.6 kIn
for oil pipelines. Depending on the size of the natural gas pipeline, the nominal
pressures are approx. 50, 64, 70 and 80 bar, respectively. Pipeline are usually buried
about one meter below ground. For most pipelines the variation of temperature is smalL
Internal corrosion depends on the material transported, but in many cases it is very low.
PWEL~ AND STORAGE TANKS 183
TABLE 6.3. Diameters of oil pipelines on the Swiss territory (after "Eidg. Rohrleitwlgsinspektoru",
Report SVDB No. 12411992)
Diameter 555.8 508.0 457.2 406.4 323.8 273.0 209 193 168.3 114.3
(mm)
Length (Ian) 121.0 3.5 0.5 25.5 70.8 11.2 4.1 10.4 0.7 1.9
184 CHAPTER 6
---
Operating Plant
•
(i)
Propane/air mixing Plant (Town gas)
Natural gas occurrence (Finsterwald)
High-pressure pipeline (1.0. > 500 mm)
High-pressure pipeline (1.0. 350 - 500 mm)
High-pressure pipeline (1.0. < 350 mm)
Local distribution network
Projected or in construction
Mechanical - - 2x 6x 2x(of
Damage which I x
(Due to Third oil outflow)
Party)
Damage to
Electrical Control
2x - - - -
Cables
Other Causes - 4x 3x - -
TABLE 6.5. Crude oil and petroleum products pipeline network
Every five years the pipelines are controlled internally with special equipment
recording any corrosion damage and wall thickness variation. At special occasions, the
186 CHAP1ER6
gas contained in a pipeline segment has to be vented to atmosphere (which takes a few
hours). A special computer program is used for calculating the venting time. Due to the
good conditions of these pipelines, which are recent, leakage problems are negligible.
TABLE 6.4 gives an indication of the small number of incidents, involving pipelines or
gas stations, that occurred in Switzerland between 1988 and 1993.
TABLE 6.6. Companies operating gas pipe lines in Switzerland
The geographic location of the pipelines and the revisions/modifications which were
made during the last years were reported in the yearly report of the Swiss Federal
Pipelines Inspectorshipl, and are mentioned in TABLE 6.5 - TABLE 6.6.
Beside the above mentioned pipelines, there are many other technical installations
which are not listed here, but include distribution stations, compressor/pumping
stations, gas and product storage, etc.
PIPELINES AND STORAGE TANKS 187
• Gasverbund Mittelland
• Gasverbund Ostschweiz
• Gas pipeline Holland-ltalien (Transit)
• Gazoduc Suisse Romand
• Erdgasleitung Ost
• Erdgasleitung West
• Erdgasleitung Rhonetal
• Erdgasleitung Zentralschweiz
• Gazoduc Orbe-Millchi (parallel pipeline to A1tavilla, 600 mm 121, in planing)
• Gazoduc Giand-Noville, VD, in planing (500 mm 121)
• Gazoduc A1tavilla - La Chaux-de Fonds - Les Verriaes
• Erdgalsleitung Finsterwald - Entlebuch
• Erdgasleitung Lichtenstein
• Erdgasleitungen der Gas-, Wasser- und Fernwllrmeversorgung der Stadt Bern
• Transitgasleitung
• Erdgasleitung Chiasso - Lugano
• Erdgasleitung Trilbbach - Chur
The risks from pipeline operation is not quite negligible as can be seen from. During the
perio11972 through 1994, according to the records taken from the data base PC-Fact;,
there were approximately 90 accidents taking place in the world, involving mostly
pressurized natural gas pipelines resulting in 516 fatalities and 1386 injuries (at the
minimum). These figures do not include Russia and Asia.
In Russia. a train induced a pipeline explosion in late '80s, causing the death of ca.
800 children. In Seoul (South Korea) a severe pipeline explosion took place in 1995.
Finally, a recent gas pipeline explosion in the USA (1994) endangered 100 people and
killed one person. Of these reported accidents, ca. 63 were explosions, the rest being
gas release, or gas release with fire. The main causes of fire and explosion were, among
others, due to: collision of tank vehicle (in one case), rupture of pipeline by bulldozer,
rupture of corroded pipeline by overstress, gas explosion in pump station. gas explosion
in building, gas leakage.
Transmission and Distribution Piping Systems (1975). There exist also several API
codes relevant to pipelines.
TABlE 6.8. Overview of pipeline accidents (including gas distribution). involving 10 or more
fatalities or injuries
Information on the failure rate of pipelines is available in the USA from a number
of sources (Figure 6.3). These include the annual reports on natural gas pipelines by the
DOT, and those on oil industry pipelines by the oil industries international study group
for conservation, CONCA WE.
The main causes of failure of pipelines in the U.S. is interference by external action.
Damages can occur from earth-moving and excavating equipment. Some major acci-
dents arising from the pipeline transport of hazardous chemicals are the fire at Hearne
(fexas, 1972); the explosion at Natchitoches (Louisiana, 1965); at Port Hudson (Mis-
souri, 1970); the toxic release at McPhearson (Kansas, 1973). These has been described
and documented elsewhere as case histories.
Pipeline Surveys
• A survey on the U.S. Safety of Interstate Natural Gas Pipelines by the Federal Power
Commission (FCP, 1966) gives information on failures in these pipelines over the period
19501965. The length of transmission pipelines covered was 204,730 miles in 1964. The
number offailures was 1058 (of which 121 compressor or metering station failures; 1115
field testing failures). Employees suffered 35 deaths and 135 injuries and non-employee 29
deaths and 87 injuries. Of the 64 deaths, 46 arose from failures in lines an 17 from failures
in compressors or metering facilities. The major source of failures were: earth-moving
equipment (279 cases); corrosion (193),· weld failure (190).
• In the year 1974, there were 335,000 miles of pipeline installed and for the period 1970-
74 there were 2093 failures (Annual Report on the Administration of the Natural Gas
Pipeline Safety Act of 1968 by the DOT). From these data the failure rate for natural gas
pipelines was 1.25 x 10-3 per mile per year.
• The annual report HSpillages from oil industry cross-country pipelines in Western
Europe H, Statistical summary of reported incidents for 1976 by CONCA WE in 1977 (9m),
states that between 1972 and 1976 the total length of oil industry pipelines increased from
15,800 km to 18,100 /em and that in the period 1972-76 there were 93 spillages. Assuming
an average of 17,700 /em pipeline over this period, the failure rate for such pipelines was
1.05 x 10-3 per km per year.
• Further information on gas pipelines operated by the British Gas Corporation (BGe) has
been reported by Knowles, Tweedle and van der Post 3 (1977), for the period 1970-77:
there had been 309 repairable incidents, of which 73 were defined as Hlost gasH incidents.
For chlorine pipeline a failure rate of 5 x 10-3 per mile per year has been estimated for
1974.
Hydrocarbon Pipelines
The majority of hydrocarbon pipelines carry flammable gases or liquids such as natural
gas, ethylene or LPG. Two hazards which may result from a break in a line carrying
flammable gas are afire or an unconfined vapor cloud explosion. If line is ruptured and
the gas escaping is ignited so that there is a flame shooting up in the air, the heat
radiated from the flame may be considerable.
In many locations the hazard is, however, slight. The greater hazard is likely to be
that of an unconfined vapor cloud explosion (UVCE) resulting from an escape of gas or
liquid from the line. This hazard is presented particularly by propane which is one of
190 CHAPTER 6
the main flammable liquids transported by pipeline. If liquid propane at 16°C escapes to
the atmosphere, about 33% flashes off as vapor and the temperature of the remaining
liquid falls below -42°C. An example of hazard survey of a hydrocarbon pipeline
(Figure 6.4) is described in "A Safety Evaluation of the Proposed St. Fergus to Moss
Morran Natural Gas Liquids and St. Fergus to Boddam Gas Pipeline", by the HSE
(Health and Safety Executive, London, 1978).
A major unconfined vapor cloud explosion occurred at Port Hudson in 1970 (750 bblof
propane (ca..l20 m3) were released by a line break and gave an explosion equivalent to
about 50 t of TNT).
Ammonia Pipelines
In the USA, farmers are using ammonia as a fertilizer and there are some very
extensive liquid anhydrous ammonia pipeline systems installed. Description of the Mid-
PIPELINES AND STORAGE TANKS 191
American pipeline System are given by Rohleder4 (1969) and by Luddekes (1975), and
of the Gulf Central Pipeline System by Inkofer6 (1969).
There is a hazard of brittle fracture, therefore the composition of the anhydrous
ammonia must be closely controlled to prevent stress corrosion cracking. The ammonia
has to contain 0.2% water. Cases where brittle fracture has propagated along a pipeline
at sound velocity are known.
Chlorine Pipelines
Pipeline transport of chlorine ;s much less extensive than that of ammonia. The chlorine
may be transported either in the vapor or in the liquid phase. The situation which might
give rise to mixed phase conditions should be avoided. The maximum operating
pressure is not to exceed 21.7 bar. Liquid chlorine has a very high coefficient of thermal
expansion and it is necessary to protect the line against hydrostatic rupture. The pipeline
may be above or below ground. Internal corrosion is very low if the chlorine is dry.
Chlorine pipelines are dealt with in Chlorine Pipelines by the Chlorine Institute
(1971, Publication No. 60). Further discussions are given by Danielson8 (1964), and by
Nelson9 (1964).
192 CHAPTER 6
In Switzerland there has never been any accident with a pipeline involving damage to
people or property. The risk arising from natural gas transport pipelines is relatively
small since in case of a small leak in a weld region, the methane gas which is lighter
than air, will escape and rise straight away in the atmosphere. In case of corrosion of the
pipe wall, or as a result of mechanical damage by earth-moving equipment, a jet may
however form (due to the high pressure of 50 to 60 bar existing in the pipeline) and
come into contact with an ignition source. Two hazards which may result from a break
in a line carrying flammable gas are a fire or an unconfined vapor cloud explosion.
The first case (fire) may happen with natural gas, whereas the second case (explosion) is
more probable with propane which may become heavier than air under special circum-
stances. In case of fire, heat radiation may be considerable. An example of pipeline risk
assessment has been described earlier in this chapter. The calculation of the conse-
quences resulting from a gas pipeline rupture may be complicated and necessitates
specialized programs. In Switzerland there are pipelines transporting natural gas, jet
kerosene/aviation gasoline, petroleum products and semi-products, however there are
no chlorine or ammonia pipelines installed.
Although Switzerland was. lucky to experience only minor incidents without conse-
quence, there has been abroad severe pipeline accidents involving natural gas explo-
sions, which took place during the last five years: Russia (late '80s), USA (1994), and
Seoul (1995). Unfortunate combination of incidents, usually related to large-scale pipe-
line fire and explosions, are causing severe accidents due to heat radiation or blast
effects.
The risk associated with accidental releases from pipelines transporting hazardous
liquids over long distances can be extremely significant Integration of risk and
reliability analysis with advanced leak detection technology through the expert system
approach, can lead to substantial risk reduction goals.
This is the main objective of a CEC-funded project in the framework of the
"Environment Programme" undertaken by a Wode Group composed of SNAM-
PROGETTI (Italy), TNO (The Netherlands), ABA (UK), and TEA.
The object of storage of liquids and gases is to smooth fluctuations in the flow in and
out a production or terminal site. By far the largest quantities of chemicals are found in
storage facilities which may be at chemical plants or in other storage terminals. In order
to satisfy the storage requirements different storage method can be considered which all
have a different economical impact, due to the different storage design. Most of the
materials held in store are flammable liquids or liquefied gases and the main concern
in the design of storage installation for such liquids is to minimize the hazard of fire.
PIPELINES AND STORAGE TANKS 193
Loss of containment from storage of toxic chemicals could in principle give rise to the
worst kind of chemical industry disaster. Much storage is located in or near urban
areas. It is generally reckoned to be relatively safe and unlikely to cause large loss of
life. Beside the risk of casualties or fatalities due to fires in a storage, the resulting
financial loss may be very high. The prevention of loss in storage is therefore very
important Good operation and maintenance depends on an effective management
system, but there is potentially a problem in situations where products such as LPG
(Liquefied propane gas) or chlorine are stored and used in non-chemical and low
technology industries (chlorine stored in public swimming pools).
The main types of storage for liquids and liquefied gases are:
1) Liquid at atmospheric pressure and temperature, so called atmospheric storage
(0.5 - 15 psig). Fixed tanks with cone (diameter up to 76 m) and domed roofs (0
up to 20 m), floating roof tank (0 up to 76 m, height up to 22 m or more).
2) Liquefied gas under pressure and at atmospheric temperature (pressure storage)
3) Liquefied gas under pressure and at low temperature (refrigerated pressure
storage, semi-refrigerated storage)
4) Liquefied gas at atmospheric pressure ( < 1 psig) and at low temperature (fully
refrigerated storage)
5) Gas under pressure (> 15 psig up to higher pressures).
Note: Since the codes mentioned are P.ritish codes, of which some are still using British
units, we indicate here the conversion factor: 1 bar = 14.5 psi; 1 m3 = 6.2898 US oil
barrels
Storage of gas under pressure is usually done in horizontal cylindrical pressure
vessels. The need for very large storage capacities for LNG (liquid natural gas) has led
to the development of pre-stressed concrete storage tanks which are protected by an
earthen embankment, or berm, and are internally insulated. Underground cavities may
also be used for storage of natural gas (also depleted oil or gas field).
Atmospheric storage tanks are quite often inerted with N2 in order to reduce the
risk of fire and explosion. The fluids so stored are referred to as volatile liquids,
flashing liquefied gas, semi-refrigerated liquefied gas, refrigerated liquefied gas and gas
under pressure.
Jumbo storage tanks ranging from 300,000 to 1,000,000 barrels capacity are
currently being used to store not only crude oil but also flammable intermediate and
finished products.
A large number of incidents occur of the bursting or collapse of atmospheric
storage tanks due to maloperation. Atmospheric pressure is maintained in a storage
tank by a vent to atmosphere. The tank may be subjected to pressure or vacuum which it
cannot withstand not only if this vent is blocked but also if it does not have the capacity
to handle the flows required to equalize the pressure. A hazard which is liable to arise as
194 CHAPTER 6
There are numerous standards, codes of practice, and guidelines which are
applicable to storage. 1O Hazard assessment studies have been also reported 11. 12. 13. 14 as
well as case histories. IS The definitions and codes indicated below were adopted from a
British transport compendium.
PIPELINES AND STORAGE TANKS 195
Filling Ratio
The maximum permissible working capacity of a storage vessel is often expressed in
terms of the filling ratio. Filling ratios are dealt with in BS 5355: 1976 Specification for
filling ratios and developed pressures for liquefied and pennanent gases, and in other
publications on storage (e.g., ICI/rasp, 1970 ISn4; LPGITA, 1974 LPG Code 1;
BCISC, 1975 Item 1).
The formula used to calculate the filling ratio is generally of the form:
e = rp / rw{l- f) (6.1)
where e is the filling ratio; rp the specific gravity of product at reference temperature; rw
the specific gravity of water at 15° C; andf, the free space (fractional). The reference
temperature is the maximum temperature which the product should reach in service.
where A is the surface area of pool (m2); C the calorific value of liquid (kJ/kg); Q
the quantity of heat radiated from the flame envelope (kW); r the liquid burning rate
(mls); p the density of liquid (kglm3); and kJ a constant equal to 0.3.
196 CHAPTER 6
The values for the liquid burning rates are given as: 0.083-0.17 mm1s. Alternatively
values of 0.1,0.3, and 0.4 inlmin are also given for 1 ft,7 ft, and 35 ft diameter pools,
respectively. For most hydrocarbon flames the heat radiation from the surface of the
cylinder (with dia. = D, height = 2D) is approximately 170-240 kW/m2. The limit for
the heat received by an adjacent storage tank is assumed to be 37.8 kW/m2.
Note: I ft = 0.3048 m; I inch = 0.0254 m.
Bunds
Some types of liquid storage tank are normally surrounded by a bund and/or provided
with a pit to retain any spillage of the liquid. Bunds are made of earth or of concrete. In
general, bunds are provided for atmospheric storage tanks and for fully refrigerated
storage tanks of liquefied gas, but not for pressure or semi-refrigerated storage of
liquefied gas or for acid or alkali storage, although this generalization needs some
qualification.
The object of buntiing is to retain the liquid so that it can be dealt with in a
controlled manner, by evaporation from a specially designed catchment evaporating
area, by foam blanketing or other measures. Bunds tend not to be used for pressure
vessels because these rarely fail, the emission when it does occur is mainly in vapor
spray form, and the dispersion of small leaks and spillage is hindered. Low walls may
be used to keep flammable liquids from some external source from reaching a storage
vessel.
PIPELINES AND STORAGE TANKS 197
For atmospheric storage tanks containing Class AlB flammable liquids a full bund
should be provided. The bund capacity should equal the total capacity of the tank. If
there is more than one tank inside the bund, the capacity should be that of the largest
tank after allowing for the displacement due to the other tanks. Low division walls
between tanks within a bund are recommended.
Some aspects of the storage of particular chemicals are considered in the next
section.
LPG Storage
LPG is stored at petroleum storage terminals, at chemical works and at many non-
chemical factories. In addition to the codes which refer explicitly to LFG (Liquefied
Flammable Gases), methods of minimizing spillage and hazards of fires can be found in
the ICI Liquefied Flammable Gases Storage and Handling Code (the ICI LFG Code)
ICIIRoSPA 1970 ISn4).
Chlorine Storage
Storage of chlorine as a liquid is mainly either in chemical works or at installations
where it is required for water treatment such as waterworks, power stations and
swimming baths. The boiling point of chlorine is 34° C. The handling and storage of
chlorine is dealt with in the BCISC Code of Practice for Chemicals with Major
Hazards: Chlorine (the Chlorine Code, 1975 Item 1), and in the various publications of
the Chlorine Institute, including Chlorine Manual (1969 Pub. 1), Facilities and oper-
ating procedures for chlorine storage (1973 Pub. 5), and other literature (1965 Pub. 57,
1970 Pub. 9, 1974 Pub. 53, 1975 Pub. 61).
Chlorine Code
The Chlorine Code gives a well defined philosophy for the storage of chlorine. In view of
the importance of the hazard of toxic release from storage the code's statement of the
problem is of interest: "liquid stored in bulk is the potential source of the most serious
incidents. Stored liquid could be released by failure of the vessel holding it, or by failure of
pipework or equipment connected to the vessel". The risk of some damage to a storage
vessel cannot be totally eliminated. the main dangers derive from failure of pipework
whether on producing or using plants or connected to storage.
Ammonia Storage
Anhydrous ammonia is stored as a liquid mainly either in chemical works or,
particularly in the USA, in storage terminals supplying agricultural users. The handling
198 CHAPIER6
of ammonia, including its storage, has been a major preoccupation of the industry and
this is reflected in the Chemical Engineering Progress series Safety in Air and
Ammonia Plants (AlChE 1960-69 Items 12-21) and Ammonia Plant Safety (and Related
Facilities; AIChE 1970-77 Items 25-32). The work of the industry Ammonia Storage
Committee has been described by W.L. Ball (1968) 17. There is a CIA Code of Practice
for the Large Scale Storage of Fully Refrigerated Anhydrous Ammonia in the United
Kingdom (the CIA Refrigerated Ammonia Storage Code) (1975 Item 8). The boiling
point of ammonia is - 33° C.
As with chlorine, the principal hazard of ammonia in storage is a large toxic
release, since a large release could give rise to a major disaster. The containment of
ammonia is therefore critical. With ammonia also the design philosophy is that the
standards should be such that vessel failure is a rarity. A review of ammonia storage
methods has been prepared by (Hale 18, 1974). The larger refrigerated atmospheric
storage tanks hold 40,000 ton.
Hydrogen Storage
Gaseous Hydrogen. The gas is stored in pressure vessels designed to the regular codes
NFC 50 A, "Gaseous Hydrogen Systems at Consumer Sites", (1973), gives separation
distances. The separation distances for a large gaseous hydrogen storage (> 15,000 scty
are 25 ft from an ignition source and 50 ft from a large flammable liquid storage ( >
1000 US gal). Pressure relief to the open air is permitted provided this is done at a safe
height.
Liquid Hydrogen. The liquefied gas is stored in pressure vessels designed to the
regular codes. Separation distances given in NFC 50 B "Liquefied Hydrogen Systl!ms at
Consumer Sites" (1973) for a large liquid hydrogen storage (> 15000 US gal) are 50 ft
from an ignition source and 100 ft from a flammable liquid storage. Again pressure
relief to the open air at a safe height is allowed. NFC 50B states that liquid hydrogen
should preferably be stored on ground higher than the other flammable liquid storage.
This, of course, is the exact opposite of the requirements for storage of products such as
LFG.
i 1 scf(standard cubic feet) = 2.832 10.2 m3, and 1 US gallon = 3.785 10- 3 m3•
PIPELINES AND STORAGE TANKS 199
Summary
The risk assessment of pipelines as well as storage tanks for gases and petroleum
products is of emerging importance for regional risk assessment and safety manage-
ment. The present chapter first surveys the transport pipelines in Switzerland: gas
pipeline networks in the content of the European environment, technical characteristics
of the gas and oil pipelines, number of incidents and damages to gas pipelines. The risk
analysis of pipelines is made within the framework of experience in the USA as well as
in Switzerland. Regarding the risk of storage tanks, the present chapter mentions
international codes of practice and regulations, and presents some methods of
calculation; storage of special products and their specific risk aspects are part of the
present analysis.
200 CHAPTER 6
References (Chapter 6)
Risk assessment for regional planing should also comprise information and statistics on
the volume and class of dangerous goods transported or stored in large quantities in the
region. This information should be stored in a dangerous goods data base which should
include:
• Information concerning the storage location, such as: location of the storage area,
type of storagelpackaging (tanks, containers, bags), tank farm arrangement, separation
distances, oil catchment equipment, etc.
• Information on the products: volume of product stored, danger class and physical
properties of the liquids or solids.
• Information for medical treatment of patients
• Information concerning the fire fighting equipment (type/number of hydrants,
typeIquantities of extinguishing product) and information regarding the availability of
firemen brigades (on a local/regional basis).
• Regional information concerning the availability and location of special protection
equipment against petroleum products, for pollution prevention and remedial actions.
Using a Geographical Information System (GIS) it is possible to visualize the
location of dangerous goods stored in large quantities in the region. An overview of the
volume and main transport flows of dangerous goods through the region might also be
represented on a synoptic map. The map should show the main transport and distribu-
tion roads through the country, and indicate the location of petrol stations along the
highways and in the cities. Additional information concerning the presenCe of highways
restaurants, the location of schools, sport stadium. hospitals, shopping centers in the
vicinity of such critical points would be useful as well. The population density along
the mains transport axes should be estimated and reported on the digital map, using a
grid of 100 x 100~. Such information is helpful in calculating the consequences of an
accident involving fire, toxic gas dispersion, or explosion, and gives also information
on the number of people to evacuate in case of an emergency.
At present such development in Switzerland is in progress and the Cantonal Planing
Offices are in the process of preparing digitized maps for use with GIS, but it will take
a few more years until the tas is accomplished. A strong coordination between the
different Cantonal Offices would be highly desirable, especially what concerns the type
202 CHAPTER 7
and format of information to be stored in regional data bases. Remote public access to
such data bases at a National level should be favored. The public should have the right
to have access to such information. as this is the case in the USA and in other countries.
Concepts for integrated network access of risk data bases and other regional data
bases of interest, using hypertext and special query languages of the fourth generation,
are being investigated in the United States and in Germany (on behalf of the European
Community).
In the framework of the PPR&S it was decided to collect and review basic
information of legal and technical matter concerning the transportation of dangerous
goods in Switzerland, as far as available. At this point it is worth noting, that there is no
law obliging trucking companies, or the Swiss railroad companies to publish any kind
of statistical information or to report publicly any kind of incident/accidents concerning
the transportation of dangerous goods. Some basic information concerning the
transportation of dangerous goods in Switzerland has been reported in Gennan
language by Kieselbach l (1993), in the framework of the PPR&S project, and selected
topics are also presented here.
It is rather difficult to obtain reliable data for use in risk assessment studies, concerning
the number of road tankers accidents, or statistics regarding the type and quantity of
dangerous goods transported through Switzerland National statistical data can be
obtained, for instan-=e, from the following source:
• Swiss Federal Office of Statistics ("Bundesamt fijr Statistik, Bern").
• "Statistisches Jahrbuch der Schweiz", an agenda with statistical data concerning
Switzerland
• Swiss Railway Company ("SBB, Schweizerische Bundesbahnen")
• Petroleum Products Association ("Erdolvereinigung")
• Fuel Dealers Association ("Brennstofthandlerverband")
• Pipeline operating companies like INTERGAS, SWISSGAS etc. have statistics
concerning products/quantities distributed by pipelines
• Swiss Federal Pipelines Inspectorship (statistics concerning pipelines incidents)
Within the framework of PPR&S we have tried to obtain from different sources
various information concerning dangerous goods transportation, e.g., from the Swiss
Petroleum Products Association, and the Swiss road tankers/truckers companies, etc.,
but without success, since no Swiss law does oblige such companies to disclaim any
information concerning the type, quantity, and destination of chemical products or other
dangerous cargo (gasoline, heating oil, industrial gases, etc.) transported by them. The
companies in question are considering such "marketing information" as confidential,
TRANSPORTATION OF DANGEROUS GOODS - GENERAL INFORMATION 203
and are not prepared to divulge them to any third party. The problems we encountered
are mention below.
a) There are very little statistics on road tankers/trucks accidents that are published in a
form useful for risk analysis. Data published abroad cannot be extrapolated, as such,
to Swiss conditions, and must be considered with prudence.
b) The quantities of goods imported in transit (by rail, road, or ship) or transported
within the country are registered at the custom-houses according to a schedule of
dues developed for fiscal purpose deserving more the preparation of financial
budgets by the Swiss government. The records cannot be used easily for risk
assessment because the quantities of various products, that do not belong to the
same class of risk, have been aggregated together.
c) Some petroleum products imported by international and diplomatic organizations, as
well as jet kerosene for the international air traffic are not subject to payment of
duty, and therefore do not appear in the statistics of the Swiss custom-houses,
respectively, of the inland revenue department.
d) The ports of Basle (on the Rhine river) represent a special case, which complicates
the statistical evaluation of the distribution of the goods traffic between ship, rail
and road. Part of the goods transported by ship are unloaded in Basle to be further
transported by rail or road for inland distribution. Record keeping by the custom-
houses does not mention the final destination of the (dangerous) goods.
e) In Switzerland we have strategic tank farm'! for petroleum products (automotive
gasoline, heating oil, jet kerosene). Imported petroleum products may be stored for
some time in such storage areas located at proximity of air ports and cities, or they
may be brought directly to the final user. Redistribution may also take place from
such storage. Thus it is not easy, without precise record keeping, to trace back the
movement of dangerous goods involving petroleum products.
t) Inland distribution of petroleum products is not always done by the oil companies,
like ESSO, SHELL, etc., but frequent use is made of private road tankers compa-
nies.
Furthermore, due to the very large autonomy of the different Cantons, several
regional statistics of interest mayor may not be made available officially to the public,
due to a new legislation on data secrecy/privacy that obliges the different Federal and
Cantonal offices to handle their data on a "confidential basis". Another difficulty arises
from the fact that the denominations of the different offices may also vary considerably
from one Cantonal Office to another, which does not simplify data inquiry.
Further statistics are available at the Cantonal level: i) - for information
concerning population density distribution, usually the Communal or Cantonal
Cadastral Office; ii) - for traffic accident statistics, the Cantonal Police Authorities,
or eventually some Cantonal or Urban Office for Statistics, or Civil Engineering Office
("Tiefbauamt"). The latter data are used for planing and taking remedial actions
pertaining to roads and traffic control.
204 CHAPI'ER 7
There are actually 10,000 products that are known to be dangerous and are being
considered by the different regulations for industry, storage, and transport. Usually the
products are classified by dangerous goods classes. According to the UNO
recommendation the products may be simply listed and numbered, or according to
RID/ARD, they have to be classified on the basis of their physical and chemical
properties and grouped in different classes according to some main properties repre-
sentative of this type of dangerous goods.
RIDIARD Regulation
The RIDIARD regulation prescribes a subdivision into 9 classes (1-9) according to
TABLE 7.1. The most important classes are the classes 3, 6.1 and 8. All the RID/ARD
classes (except the classes 2 and 7) are further subdivided into three subdivisions giving
an indication of the acuity of the danger (Le., a- extremely dangerous, b- dangerous, c-
less dangerous).
Packaging Groups/Codes
In the regulations we find, according to the same scheme, the packaging groups I-llI
corresponding to the danger class (a-c). The danger classes with their corresponding
packaging groups and code a presented in TABLE 7.2.
The significance of this packaging code is that the packaging has to comply with
some min~mal exigency corresponding to the danger class. Transport of dangerous
goods by post is principally forbidden. According to international regulations, it is only
allowed to transport dangerous goods internationally, when these are properly packed in
accordance with the packaging groups and codes corresponding to the representative
danger class. The definition of the tanks and containers to be used in conjunction with
TRANSPORTATION OF DANGEROUS GOODS - GENERAL INFORMATION 205
7 Radioactive productsy
8 Caustic products
TABLE 7.2. Danger classes with their corresponding packaging groups and codes
RIDIARD have been described in detail by Kieselbach 1 and are not reproduced here.
The dangerous goods to be transported must be marked accordingly (following the main
classification scheme). The road tankers loaded with dangerous goods must be provided
with an orange colored label. The danger class is defined by a number located in the
upper part of this label and a second number (so-called UNO number) in the lower part
gives an indication of the content of the tank.
206 CHAPTER 7
lifM.: the tonnage in the folWwing tables is upressed either in million tOIU
(Mt). or tlroustJ1ld toIlS (let). or toIlS (t).
The Swiss Federal Railways are not only transporting goods by train, but offer a
special customer service, SBB-CARGO, for delivering the goods by truck to the
customers' place.
According to an information published by SBB-CARGO (111993), 50 tons of goods
were transported in transit or inland, of which 7 tons were dangerous goods. The period
was not mentioned but this statistic probably refers to the year 1992 or possibly 1991.
The distribution of the dangerous goods transports between rail and road is shown in
TABLE 7.4.
TABLE 7.4. Distribution of dangerous goods' transports between rail and road
(according to SBB-CARGO. 1/1993)
Extremely Dangerous 43 57
Less Dangerous 38 62
TABLE 7.6. Rail transport statistics by product category (according to Swiss Federal Office
of Transportation; BAV, 1990)
The figures of TABLE 7.6 have been reported by the Swiss Federal Office of
Transportation (BAV, Schweizerisches Bundesamt fur Verkehr). The problem with this
kind of information reporting is that the content of the category/class is not defmed at
all and does not correspond to the definitions uSf_i by RID/ARD for the classification of
dangerous goods. The United Nations Organization, UNO, has published in 1989 a
statistic regarding the transportation of different kinds of goods. The data concerning
Switzerland are reproduced in TABLE 7.7.
TABLE 7.7. Dangerous goods transports in Switzerland (according to UNO-statistics)
Another statistic compiled in 1987 comes from the Swiss Society of the Chemical
Industry (SOCI, "Schweizerische Oesellschaft der chemischen Industrie") and was
established on the basis of a questionnaire which was send to 28 member companies
(TABLE 7.8).
The distribution of the SOCI data according to RID/ARD categories is shown in
TABLE 7.9.
TABLE 7.9. Statistics on dangerous goods' distribution accord-
ing to dangerous goods classes (Survey established by SGCI. the
Swiss Society of the Chemical Industry ) in 1987
1 1'078.0 0.2
2 25'437.5 3.6
3 99'149.7 14.2
4.1 21'197.0 3.5
4.2 0.5 0.0
4.3 0.5 0.0
5.1 6'063.0 0.9
5.2 4'700.0 0.7
6.1 55'027.0 7.9
7 0.0 0.0
8 483'889.0 69.1
The SOCI statistics present a difference of 400 t. It seems probable that the transport
of petroleum products has been omitted or reporting was incomplete.
TRANSPORTATION OF DANGEROUS GOODS - GENERAL INFORMATION 209
SGCI has also reported a change of the Swiss national economic index for the period
1987 to 1990 (TABLE 7.10), and it seems reasonable to admit that the quantities have
also increased proportionally to this index change.
TABLE 7.11 show an approximate distribution of the goods volume between ship
and rail, based on data published for Germany and extrapolated to Switzerland
conditions. Within Switzerland, the goods are transported only by road or rail, so that
we can assume that goods imported by sbiplbarge are to be distributed on both transport
medium. We have to consider two transportation scheme:
• Sea vessel - Sbiplbarge on the Rhine - RaiVroad distribution
• Shiplbarge on inland waterwaysllake - RaiVroad distribution
The proportion of the ship freight to be further allocated to the rail transport must be
guessed (e.g., 40-60%) and added to the amount mentioned for rail (in the second
column). The results shows up in the third to fith columns:
TABLE 7.11. Distribution of the transported tonnage between different transport medium. (Extra-
polation from German statistics to Swiss conditions, assuming a rail share of 40, SO, and 60%)
Transport Trame Swiss Share (%) Swiss Share (%) Swiss Shan' (%)
Medium Share Rail =40% or Rail = 50% or Rail = 60% cf
(%) Ship freight Ship freight Ship freight
TABLE 7.12 shows the traffic volume resulting from different rail share
assumptions (40, 50, and 60%). If we assume a 50% share (i.e., an equal distribution
between rail and road). The corresponding distribution with respect to goods classes is
shown in TABLE 7.13.
There is another possibility of estimating the distribution of goods transported by rail
and road in Switzerland which is based on the known quantities of products belonging
to "Class 3", which were transported to the end consumer (TABLE 7.13). It was
assumed that 60% of the goods are transported by road and 40% by rail. The
extrapolation based on "Class 3" results in too low proportion of dangerous goods
210 CHAPTER 7
transported by rail, which means that there must exist a different percentage distribution
between the dangerous goods.
TABLE 7.12. Distribution of dangerous products between rail and road (extrapolation)
Product Ra il Road
(acc:ording to BAV (estimate)
1990)
Chlorine. Within the present study it was not possible to obtain independent
information concerning the production of chlorine in Switzerland. An estimation made
by Green Peace on the basis of insider information indicates the quantities shown in
TABLE 7.14. Transport of chlorine in Switzerland is taking place in liquid form in
pressurized tankers of 22'000 I or 50'000 I content, mobile tank of 1'000 1 or in cylinders
of 50 I content.
Liquid Propane Gas. LPG for sale is actually a mixture of propane-butane (so-called
C-mixture) which is traded under the name "propane". This mixture which contains
TRANSPORTATION OF DANGEROUS GOODS - GENERAL INFORMATION 211
97% propane behaves physically like pure propane. The total quantities of LPG
transported in Switzerland can be derived from the statistics of the Swiss Petroleum
Products' Association (EV, "Erdolvereinigung"). LPG is transported in rail-tanker of
105'000 I content and road tankers 35'000 I content, in tank-containers with a content of
20'000 I and in gas bottles of 2 to 60 1. The rail-tankers usually build a continuous
block of single cars transporting the gas directly from the refmery to the customer'
place.
TABLE 7.14. Estimated chlorine production in Switzerland
Vinyl Chloride. Vinyl chloride is used as raw product for the fabrication of PVC-
plastic. It is imported as special liquefied gas with rail-tankers of 70'000 I exclusively,
assembled in blocks, some of which are also transiting through Switzerland.
Air-mixtures. These include oxygen, nitrogen, and argon gases pure or in mixture with
air. These gases are either imported or produced d.irect!y in Switzerland. The companies
trading w~th these gases are usually daughter compames of foreign holdings, which are
also member of INTERGAS, located in Bern. The companies active in Switzerland are
represented inTABLE 7.15.
It was not possible to obtain any information on type and quantities of gases
transported from these companies. Transport is taking place in liquid form by
refrigerated road and rail tankers of 60'000 I, and as compressed gas in cylinders of 50
l. Such gases are inert and do not present any risk during transportation.
The distribution of the imported petroleum products between the different transport
medium is shown in TABLE 7.17. The transport capacity of the firms belonging to the
fuel dealers as well as the type of road tankers are indicated in TABLE 7.18. The
TRANSPORTATION OF DANGEROUS GOODS - GENERAL INFORMATION 213
market share of members of the Petroleum Products' Association have been estimated
and are shown in TABLE 7.19.
TABLE 7.17. Distribution of imported pettoleum products between the
different transport medium (1991)
TABLE 7.18. Transport capacity of the fl11l1S belonging to the Association of Fuel Dealers
TABLE 7.19. Market share (%) of the members of the Pettoleum Products
Association (Estimate)
TABLE 7.20. Distribution of the pettoleum products (in 1,000 tons) according to the
different Cantons
TABLE 7.21. Geographic dislribution of petroleum products imports (according to Swiss Petroleum
Association, 1991).
Petroleum Products
(1,000 Ions)
MUlhausen 11 - - - · · · - -
Schaffhausen 28 - - - - · · · -
Kreuzlingen ~ - - · · - · - -
Bern 41 51 - - - - - - .
Luzem 24 30 - - - - - - -
A1tdorf 1 3 - - - - - - .
North-West 261 23 . - 2 - 510 242 1038
Val10rbe 47 46 - - - - · - 93
Fribourg 40 20 - - - - - - 60
AigIe 40 84 - - 52 30 - · 206
Tessin 24 23 - - - · 64 12 123
Poschiavo - . - - - - 2 - 2
Simplon 2 - - - - - - - 2
East 87 20 - - - - 71 ~ 267
WiVSG 22 40 - - - - - - 62
Sennwald 50 - - - - - - · 50
Erlen 4 11 - - - · · - 15
Landquart 3 18 - - - - - - 21
Bund 5 3 - - - - - - 8
TOTAL 2199 951 1006 453 585 139 933 1414 7659
Note: The rubrique H.OiI (Heatmg Oil) and DIesel also UlCludes kerosene.
There are actually four main petroleum and products pipelines in operation in
Switzerland (see also Chapter 6). These are:
216 CHAPIER 7
The statistics presented in this section were prepared by the Swiss Petroleum
Association, EV ("Erdolvereinigung"). The original report covers, however, a larger
period than this is the case for the data presented here.
Swiss Market. The. sales of petroleum products in Switzerland (TABLE 7.23) totalized
12,732 million tons for the year 1994, of which motor fuel 47.4%, heating oil 45.6%,
and other products 7.0% (including refmery own consumption).
TRANSPORTATION OF DANGEROUS GOODS - GENERAL INFORMATION 217
~: 'Without the products not liable to custom duty, such as Samnau, TEE, Diplomatic and
international organizations, and without the jet kerosine for the international air traffic (not liable to duty),
and without the importations which do not neccessitale dispatching way-bills.
Petroleum Products in Transit by Pipelines. The pipeline connecting the area of Genua
!Marseille to Switzerland transported a total of 5,890 million tons crude oil, spikes
(semi-products), and finished products into Switzerland (including finished products
not liable to custom duty).
The pipeline connecting Central Europe to Italy transited 7,450 million tons crude
oil and finished products through the Canton Graubilnden. TABLE 7.27 shows the
tonnage of petroleum products transported by pipelines in Switzerland, including toll-
free products, aviation gasoline, and jet kerosine.
TRANSPORTATION OF DANGEROUS GOODS - GENERAL INFORMATION 219
During the year 1994 there was no damage or incident to the pipeline network
covering the Swiss territory.
TABLE 7.27. Quantities ofPeIroleum products transported by pipelines (year 1994)
Ship Transports on the Rhine River. The petroleum products imported to Switzerland
by ship transport on the Rhine river (TABLE 7.28), amounted to 4,161 million tons for
the year 1994 (inclunding products not liable to custom duty).
TABLE 7.28. Petroleum products imported to Switzerland by ship
Product Year 1983 Year 1991 Year 1992 Year 1993 Year 1994
% % % % %
With a share of 56.1% (54.7% in 1993) of the total goods transported by ship, the
petroleum products imports represent quantitatively the largest amount of the products
arriving at the ports of both Basel (located on Swiss and German territory as well). The
export by ship of petroleum products (mainly heavy heating oil) represented 48.6% of
TRANSPORTATION OF DANGEROUS GOODS - GENERAL INFORMATION 221
the ship transports in the Basel area. The distribution of petroleum product sales by
categories is given in TABLE 7.30.
Railroad Transports. Inland railroad transports of petroleum products from the ports
of Basel amounted to 28% (1,165.08 million tons) of the total quantities imported by
ship for the year 1994 (against 34% in 1993).
Service Stations. At the end of the year 1994 the number of service stations delivering
automotive petrol and diesel oil amounted in Switzerland to 3841 fuel stations. The
sales of the self-service fuel stations represented 89.8% of the total sales of petrol and
diesel oil, on average, 1,295 million liters per station. The fuel stations along the
national highways sold, on average, 5,560 million liters per station.
In Switzerland the reporting of accidents by the local police or the highway police
should definitively be done according to an accident protocol, which still remains to be
dermed and agreed on. Nowadays both statistical data concerning volume and type of
dangerous goods in transit and the number, type and consequences of transport accident
are not easily available. The statistics do not allow conclusions on accident rate per
dangerous goods category to be drawn. It is also not possible to establish a flow chart of
the chemicals and other dangerous goods transported within or through Switzerland
There is a necessity for implementing a data base concerned with dangerous goods
transport incidents, as this is done in the U.s. or in other countries. The actual
legislation does not oblige the transport firms and trucking companies to announce
dangerous transports to the authorities nor are they at present obliged to supply
information on goods movements and volume transported
The reporting of accidents should also attempt to break the causes of an accident
down into two main categories, human error or mechanical failure. Among the types of
accidents we find rollover, leak, spills, fire, (explosion), equipment failure. Rollover
and spills are usually quite frequent. Liquids transported in bulk require special driver
skills because of the high center of gravity and liquid movements which can easily
cause rollover.
222 CHAPI'ER 7
The Public Safety Service for the control of dangerous goods transports (Cantonal
Administration! Cantonal Police) should ensure that:
• Material safety data sheets for all products are carried in the cabin of the trucks.
• Detailed maps of all consignee loading and unloading areas and emergency
equipment location are Available at the Public Safety Service.
The Public Safety Service should also assist in the development of the trucking
companies' policy and procedure manual. Vital elements in preventing motor vehicle
accidents are the driver's skill and knowledge, both of which can be improved through
appropriate training. Driver attitude is affected by the company's philosophy. In view of
increasing concurrence in the transportation business, the pressure on the drivers from
the side of the companies is increasing. As a result violation of speed limits, rest period,
etc. are a frequent cause of accident due to fatigue, speed excess or inattention.
TABIE 7.32. Transport of dangerous goods (TOO) in Western Germany. and related
accident statistics3
TDG toxicto
water
(average for 1982-
1991)
Number of 24.0 306.0 82.0 412.0
Accidents
Volume released 137.0 411.0 386.0 934.0
to environment
(m3)
Risk Index
Accidental Risk 2.8 18.0 6.9 -
(A.R./Mrd t-km)
Index-value: 1 6.5 2.5 -
Rail =I
Damage Risk 15.8 24.2 32.5 -
(m3/Mrd ton-km)
Index-value: 1 1.5 2.1 -
Rail =I
The share of road-trucks accidents is 74%. Regarding the accidents of road tankers
and ships, 85% of the total load was released to the environment. The volume released
to environment per accident was the lowest for road transport (1.3 m3). The quantities
involved in ship and rail accidents were much larger (4.7 m3 per ship accident and 5.7
m3 per rail accident). A comparison of accidents, respectively volume released to
environment based on the transport yield in Mrd ton-kID, one obtains two risk indices,
showing that rail transportation is more secure than the other two.
The transport of dangerous goods by road (TABLE 7.31) resulted in 113 accidents
involving only road tankers for the period 1975-1976,4 and 248 dangerous goods
transports for the period 1982-1984. 5 Reports differentiating the accident according to
parcel transport and road tanker transport were also reported for different periods:
• Period 1987-1989 (reported in 1990): Accidents involving parcel transporta-
tion6
• Period 1990 (Reported in 1991): Accidents involving parcel transportation7
• Period 1988-1990 (reported in 1992): Accidents involving water toxic pro-
ducts. 8
• Period 1989-1991 (reported in 1993): Accidents involving road tankers9.
224 CHAPTER 7
The differences between the results of the different surveys under consideration are
due to different definitions of the accidents, transports, and product involved. It was
also established that not every accident is recorded in the same way, nor that every
accident is recorded at all by the different statistical services in charge of producing
official surveys. For the period 1989-1991, there were 131 accidents recorded
(involving product release or damaged valves/piping). The accident frequency has
dropped as compared to the year 1980, but the number of participants involved in an
accident bas remained nearly constant
TABLE 7.33 shows the distribution of accidents in West-Germany according to the
location where the accident occured (i.e., inside a locality, on country roads, on high-
ways).
TABlE 7.33. Accident distribution in West..<Jermany for the period 1989·1991
Some considerations were also given to the accident circumstances and these are
summarized below:
Day and Time. 120 of 131 accidents occurred from Monday to Friday and 101 of 131
happened at day-time.
Type of Accidents. 60 of 131 accidents were due to driver's own fault (of which 53%
were &elf-accidents), 33 of 131 were due to side collisions or due to front-to-rear
collisio.ls (ca. 50%).
Kind of Accident. It includes deviation of the vehicle to the left side or right side of
the roadllane (70 of 131 accidents). The next most common cause was driving onto the
next vehicle (front-end collision type). Usually this kind of accident occurred on
Highway, and was occasioned by a foreign vehicle colliding into the rear of a road-
tanker.
Responsibility. The driver responsible for the accident was, in most cases, the tanker-
driver (70%). This is due to the high rate of self-accidents (53%). In the case of
accidents involving two or more vehicles the tanker-driver was only found responsible
in one third of the cases.
smaller on country roads, and the least severe in localities. In 74 accidents involving
dangerous goods, there were altogether 166 people injured or killed (table below):
TABLE 7.34. Damage to people resulting from accidents involving the transportation
of dangerous goods
The dangerous goods, as such, were the direct cause of two people's fatality and of
one person being injured (TABLE 7.34).
Regarding the distribution of the transported goods according to the RID danger
classes (TABLE 7.35), 103 (82%) of the 126 trucks had an accident involving danger-
ous goods of class 3 (flammable products).
TABLE 7.35. Accidents classified by location and dangerous goods class (RID)
Concerning the number of accidents involving "Danger Class 3", 38% of them
happened on highways, 48% on country roads and only 14% in localities. A release of
the tanker content was involved in 118 of 153 accidents. It was reported that the tank-
wall was damaged in 65 cases. On average the accidents involved 16'000 1 or 3'SOO kg
of dangerous goods. The quantity released to the environment (on average 4'800 liters
of which 1'000 1 could be recuperated) is larger for accidents happening on highways
than in localities or on country roads. For the period 1989-1991, there were six severe
accidents involving more than 30'000 liters lost to the environment. The costs of such
accident have been estimated and reported by "Bundesanstalt fUr Strassenwesen,
Wiesbaden".10
226 CHAPTER 7
France
French statistics on trucks accident have been reported by Degranges ll (1989). Data on
risk aSsessment for anhydrous ammonia have been compiled by Castellano et al.,12
(1989), and propane transportation by Hubert, et al.,13 (1990). study for the region of
Lyon has been presented by Hubert, et al.,14 (1988). The French Ministry of Transports
is publishing yearly accident statistics. is Statistics for the French Railways, SNCF, are
also available from the Ministry of Transports in Paris.
Great-Britain
Transport accident frequency data have been reported by Appleton,16 (1988). The
Health and Safety Commission is also a source of interesting publications on
transportation. 17
USA
The following reports are available on: Severity of transportation accidents involving
large packages in the U.S. (Dennis et al.,18 1978), and risk assessment for transporting
propane by truck and rail (Geffen, et al.,19 1980).
Holland
A study regarding LPG transport in the Netherlands has been published by TN020
(1983).
vehicles, tanks, containers and packaging. Another concern is the design of better road
and ancillary equipment, such as oil catchment canalization, barriers, improved
signaling and speed control system, etc.
In Switzerland the reporting of accidents by the local police or the highway police
need to be done according to an accident protocol, which still needs to be defined and
agreed on. Nowadays both statistical data concerning volume and type of dangerous
goods in transit and the number, type and consequences of transport accident are not
easily available. The statistics do not allow conclusions on accident rate per dangerous
goods category to be drawn. It is also not possible to establish a flow chart of the
chemicals and other dangerous goods transported within or through Switzerland. There
is a necessity for implementing a data base concerned with dangerous goods transport
incideQts, as this is done in the U.S., or in other countries. The actual legislation does
not oblige the transport firms and trucking companies to announce dangerous transports
to the authorities nor are they at present obliged to supply information on goods
movements and volume.
. The reporting of accidents should also attempt to break the causes of an accident
down into two main categories, human error or mechanical failure. Among the types of
accidents we find rollover, leak, spills, fire, (explosion), equipment failure. Rollover
and spills are usually quite frequent Liquids transported in bulk require special driver
skills because of the high center of gravity and liquid movements which can easily
cause rollover.
Public Safety Service for the control of dangerous goods transports (Cantonal
Administration / Cantonal Police) should ensure that:
• material safety data sheets for all products are carried in the cabin of the trucks.
• detailed maps of all consignee loading and unloading areas and emergency
equipment location are available at the Public safety Service.
The Public Safety Service should also assist in the development of the trucking
company's policy and procedure manual. Vital elements in preventing motor vehicle
accidents are the driver's skill and knowledge, both of which can be improved through
appropriate training. Driver attitude is affected by the company's philosophy. In view of
increasing concurrence in the transportation business, the pressure on the drivers from
the side of the companies is increasing. As a result violation of speed limits, rest period,
etc., are a frequent cause of accident due to fatigue, speed excess or inattention.
The equipment used for inland road transportation of dangerous goods may vary and
the carriers are relatively free to use a wide range of containers and vehicles, given such
restrictions as physical state of the product, and need for pressurization or temperature
TRANSPORTATION OF DANGEROUS GOODS - GENERAL INFORMATION 229
dangerous goods, can only be shipped across the border in compliance with the rules
stated by such regulations and ordinances. Moreover the maximum allowable quantities
and packaging type, which are prescribed, must be adhered to.
The prescriptions concerning construction, testing and examination, permit for use
for dangerous goods' transportation are defined in the following regulations:
• The UNO recommendations are aiming at harmonizing the national regulations in
order to assure the same quality standards for all tanks and containers used for
transportation.
• Both the IMDG code and the new edition of the U.S. Code of Federal Regulations
have taken over the recommendations of the UNO concerning ship transport. The
ADRIRID and SDRIRSD regulations, as well as other European regUlations, however
are based mainly on the 100 years older RID code. Differences between the inter-
national and national codes are to be found regarding the wall thickness of containers
and tanks. The problematic has been explained in an EMPA-report1 (1993).
What concerns the safety of tanks in case of accidents, there has been several test
series conducted by specialized institutions like for instance the EMPA (Swiss Federal
Laboratories for Materials Testing and Research). Three kinds of experiments
concerning the resistance of tanks against piercing have been published:
• Experiments with plates or shells on testing machines
• Experiments with tanks on a full scale
• Experiments with models
The results of the experiments have shown that real accidents are difficult to
reproduce in an experimental set up. Tht: following conclusions can be drawn:
- road tankers with aluminum container, may burst as a result of a chock, when the
tank is full. The resistance to perforation is smaller than for steel containers.
regarding rail tankers, the nonnal containers have a wall thickness of 6 mm, and
the containers for transporting liquefied gases have a wall thickness of 12 to 15
mm. In the latter case a perforation of the container wall is very improbable.
- all the tanks are designed to withstand the internal vapor pressure at 50°C.
Infonnation concerning type and number of rail and road tankers used in Switzer-
land (TABLE 7.39) can be obtained from the following sources: .
• Federal Office for Statistics (for road tankers).
• Data base of EGI ("Eidg. GeJahrgutinspektorat"), that is the Swiss Association
in charge of Controlling Pressure Vessels (SVDB, "Schweizerischer Verein for
Druckbehiilteriiberwachung").
• Data base of ZfW ("Zugforderungs- und Werkstiittendienst SBB"), that is the
repair workshops services of SBB. This data base maintains records concerning
the type and number of rail tankers in service.
230 CHAPTER 7
The next set of tables indicates the number of rail tankers with a given wall
thickness (TABLE 7.40), and the utilization of rail tankers for liquefied gases for the
year 1988 (TABLE 7.41). The typical dimensions ofraillroad tankers and containers are
given in TABLE 7.42.
TABLE 7.40. Wall thickness of rail tankers used for petroleum produ·::ts
Ammonia 4
C-Mixture (Propane) 80
C(h 6
Ethylenoxide 2
Chlorine 32
Sulfur dioxide 6
(h,N2, TK 13
Bromine 3
Phosphorus trichloride 5
Methylene chloride 3
Vinyl-chloride 6
Butadien 2
Dimethyl-sulfate 2
Di-chloride-fluoro- I
methane
At the present time there are no reliable tanks and components failure probability data
available for Switzerland. Based on data published in the open literature, probability of
material releases were derived for Swiss conditions (Kieselbach 1, 1993). Such data
(TABLE 7.43) must be used with caution since the conditions prevailing in the
countries of origin are quite different from the Swiss conditions:
232 CHAPTER?
• The road conditions are different due to the presence of mountains and hills. The
maintenance is usually better, but the road are subjected to variable climatic
conditions like fog, snow, and ice in winter.
• The driving habits and regulations are not the same: restrictions regarding
driving and rest periods, no truck driving at night, lower average speed, lower
transported weight, etc.
If we assume, for Switzerland, 5'000 road tankers carrying petrol and 100'000 km
driven per tanker and per year, we obtain 5 x 10.8 truck-km per year. This means that
we should reckon with at least one accident (of any type, except rear-end collision).
But, indeed reality looks far better.
TABLE 7.43. Probability of release per road tanker-krn
tanks on the basis of event trees. However it must be stated that there are not many data
from foreign studies that may be transposed to Swiss conditions and particularities.
The principles underlying the design and operation of tanks and vessels for the
containment of hazardous substances are broadly similar for transport by road, rail
and sea. Relevant codes are: the UN Transport Code, the IMCO Code ('Intergovern-
mental Maritime Consultative Organization), RID, ADR and the (British) Home Office
Containers Report.
Reloading points for road and rail tankers are areas in which a large number of
accidents occur. There are three broad categories of hazard for a tanker terminal:
• Fire/explosion of product in the tanker
• Fire and explosion from leaks and spillage
• Accident arising from transfer of the wrong products.
Terminals are discussed in Truck Loading Rack Safety by the alA (Oil Insurance
Association, USA; Item 8), in API Bull. 1003 (1975, Item 29) and by Hughes?3
Relevant codes are NFC 30 Flammable and Combustible Liquids Code (1973); and
NFC 58 Liquefied Petroleum Gases, Storage and Handling (1974); and the ICI LFG
Code.
Summary
Risk assessment for regional planing comprises information and statistics of the volume
and class of dangerous goods transported or stored in large quantities in a given region.
Specialized databases and information procedures were lately developed in view of
assisting the risk analysis calculation as well as comparative assessment of various
transportation systems or technologies involved in the given technical process. The
present chapter introduces to the reader information, sources of statistical data as well as
the type and quantity of goods transported in Switzerland. Importation / exportation
statistics concerning petroleum products are presented in detail. Traffic accident
statistics involving hazardous materials are presented for Switzerland. Similar data are
given for Germany. Statistical information on rail/road transport containers and the
reliability of tank/components failure are also included in this chapter.
234 CHAP'IER 7
References (Chapter 7)
The data basesi concerning dangerous goods and health hazards, listed in TABLE S.l,
provide useful information for risk assessment studies.
TABLE 8.1. Data bases for hazardous chemicals and materials
i Note: Ai; listed in CD-ROM in print (1988-1989), An International Guide, Meckler Publishing House.
236 CHAPTERS
The list is not exhaustive and there exist, however, other data bases which belong to
private organizations and are not always easily accessible to the public (e.g., transport
and distribution of petroleum products; accident statistics of trucking companies, etc.).
Paul Scherrer Institute (PSI) and the Swiss Federal Institute of Technology in ZUrich
(ETHZ) have established in 1993 a long-teno project) on "Comprehensive Assessment
of Energy Systems." The ultimate goal of this project is an integrated evaluation,
covering the risk-related, environmental and economical aspects associated with
different energy systems of interest for Swiss conditions. Within the framework of this
project a "PSI-Severe Accident D(ltabase" was created, which includes valuable
infonnation from other data-bases and organization, with emphasis on energy-related
events. TABLE S.2 indicates the origin of the data which were used for, or added to the
PSI data base.
TABLE 8.2. Data bases containing historical accident data
The system is also capable of integration into international data bases such as the
UK Health and Safety Executive I Safety Reliability Directorate "Major Incident Data
DATA BASES AND COMPUTER SUPPORT FOR RISK ASSESSMENT 237
Service" (MIDDAS). The AFIRS method of coding and the Australian Standard "AS
2577-1983" and certain fields from MHIDAS were applied.
The reports were entered into a DBase-ill plus and data analysis was carried out using
Lotus 1-2-3. Data were collected on the main hazardous materials by United Nations
(UN) Number. A total of 106 UN Number was reported, and the percentage of incidents
is shown below:
Energy Library
Type CD-ROM Database - bibliographic reference, updated quarterly.
Computer: IBM PC, XT, AT, PS/2 or compatible: Apple Macintosh.
Available : from OCLC USA.
Description: This data base provides users with information on oil, coal, nuclear, solar
and other related energy sources. Over 290,000 records on the disc deal with hundreds
of energy related topics.
ENFLEXINFO
Type : CD-ROM Database (full text), updated by-monthly. There is a data base
full text demo disc available.
Available: from ERM Computer Services Inc. - USA
Description: Provides current compilation and the full text of US Federal and state
environmental regulations. The regulations are indexed in six ways; reference applica-
bility, requirement, full text and data.
Environment Library
Type CD-ROM Database - bibliographic reference, updated quarterly.
Computer: IBMPC, XT, AT, PS/2 or compatible: Apple Macintosh.
Available : from silver Platter - Worldwide Distribution Network.
Description: This data base provides over 400,000 records on environment, both
historical and current, on all topics dealing with pollution and environmental issues of
concern to environmental researchers.
HA7ARD-MASTER
Type CD-ROM Full text, reference.
Computer: IBM PC, PS/2 or compatible.
Available: from USA-Technical Services Associates Inc.
Description: This disc offers a comprehensive library of safety transportation, and
environmental regulations with a powerful, accurate, and easy to use search and
240 CHAPTERS
PC-FACTS
Description: PC-Facts is a chemical events data base developed by TNO Institute in
Apeldoom (Holland). The system has now been acquired by the Swiss Military Atomic
and Chemical Laboratory ("AC-Laboratorium") and is located in Spiez (Switzerland).
The information stored in PC-FACTS can be used to prepare reports on risk 1IDalysis
based on accident/incidents reported. i
POISINDEX
Type : CD-ROM Database - bibliographic reference, updated quarterly.
Computer: mMPC XT, AT or 100% compatible.
Available: from Micromedex Inc. (USA) Microinfo Ltd. (UK).
Description: This is a detailed toxicology data base designed to identify and provide
ingredient information for over one-half million commercial, industrial pharmaceutical,
zoological and, botanical substances. Also provides detailed symptomatology and
management/treatment protocols in the event of a toxicology problem due to ingestion
dermal absorption, eye exposure or inhalation of any of the substances listed. The data
base may be searched by brand name manufacturers name, generic or chemical name,
sheet or slang terminology or botanical and common name.
those on the, "extremely hazardous substances" list. Agents are indexed by chemical
names, synonyms and commonly associated numbers.
~lle Transponation Statistics Annual Repon is a publication that summarizes the state
o~ the Nation's transportation systems and the issues acd consequences of maintaining
such a diverse and complex network, the quality of statistics used to characterize the
transportation system, and planned efforts by the BTS to improve the quality of
statistics. All four transportation modes - airways, highways, railways, and waterways -
are examined through available data and statistical studies. The 1995 edition of the
TSAR will be available in July 1995 and will focus on transportation and the economy.
In addition, the 1995 TSAR will provide analysis and interpretation of the information
contained in the National Transportation Statistics cited above.
illustrated. The 1995 NTS has been redesigned to complement the Transportation
Statistics Annual Report cited below. The NTS is also available on diskette in Excel and
Lotus formats.
The FAA Statistical Handbook of Aviation is published annually by the Federal
Aviation Administration (FAA). Its prime purpose is to server as a convenient source
for historical data and to assist in evaluating progress. The Handbook should provide a
valuable source of information for organizations interested in aviation.
The 1993 Commodity Flow Survey (CFS) is the first comprehensive effort since
1977 to learn where and how goods are shipped in the U.S. The Census Bureau has
released preliminary national tabulations of value, tons, and ton miles of shipments by
mode of transportation, commodity, distance shipped, and size of shipment. In addition,
the DOT and the Census Bureau have released more detailed information from the CFS
including the flows of individual commodity groups between and within States, as well
as flows among 89 National Transportation Analysis Regions.
exists in the areas of reprocessing, fuel fabrication, waste management, tritium, and the
Savannah River Technology Center. They provide data for use such as failure rate
analyses, equipment reliability and breakdown, studies, etc.
CEAM serve~ as the focal point for Office of Research and Development's multimedia
exposure assessment modeling and ecological risk assessment. CEAM information and
software products are available through the World Wide Web (WWW).
The information and software available through WWW were pr.::pared in support of
U.S. EPA Sdentific Programming Contract, by the Exposure R~search Laboratory,
Athens, GA, USA.
A transportation risk management system (XTRIM) is available for road and rail
transport in Denmark and the Lombardia (Italy). The underlying models for accident
frequencies and consequences are mainly based on empirical analysis of historical
accident descriptions.
A study on the risks of transporting dangerous goods in inland waterways, restricted
waters and coastal waters is currently being undertaken. An accident datallase compris-
ing 3521 accidents happened during transport of dangerous goods has been established
using more than 20 different sources from the open literature. 1869 of the accident
records describe marine and inland waterways accidents. Each accident record contains
42 information fields. The database is operated under Oracle.
An important contact has been made to Ministry of Transport and Public Works in
Holland, where they have a large knowledge on inland waterways transportation. It is
hoped to receive 17,000-18,000 accident descriptions from a ten year period on the
DATA BASES AND COMPUTER SUPPORT FOR RISK ASSESSMENT 245
inland waterways of Holland. The future plan for the transportation risk research is to
develop frequency and consequence models suitable for the XTRIM enabling this to
make risk estimates for road, rail, marine and inland waterways transport of dangerous
goods within Europe.
8.6. Spatial Data and Geographic Information Systems Within the U.S. Bureau of
Transportation Statistics (BTS)
Overview
BTS is compiling a geographic data!Jase to support research, analysis, and decision
making across all modes of transportation. The database is designed primarily to meet
the needs of DOT at the national level, but will have major applications at State and
local scales throughout the transportation community. Most of the database is to be
published as a National, Transportation Atlas, including base maps that portray the
extent and connectivity of the nation's, transportation system. thematic maps' that
portray the use, and consequences of transportation, and electronic files that are used to
develop the maps and analyze transportation networks. The remaining part of the
database will be used for internal BTS analyses, but will be left to others to publish.
The National Transportation Atlas database is organized into four layers of inform-
ation:
• The facilities layer represents the location, physical characteristics, and connectivity
of highways, railroads, waterways, fixed guideway transit, airports, pipelines, terminals.
bridges, locks, structures.
• The service layer represents transportation services on and across the transportation
facilities, such as bus lines and railroad trackage rights.
I A thematic map is a map that displays one or more specific types of information usually using symbols to
represent different classes or movements of information.
246 CHAPTER 8
Attributes. Two types of attributes have been defmed for the National Transportation
Atlas databases: locational information, and transportation attributes.
Locational attributes provide information necessary to represent the data spatially,
to uniquely identify every element in the databases, to provide connectivity of trans-
portation facilities and services, and to provide referencing systems for transportation
attributes. In general, these attributes will be located in a single database.; The proposed
attributes include facility network link, facility network node, facility terminal and other
structures, service network, flow layer point, line, boundary, and area databases,
respectively. Facility network nodes and facility terminals are located in separate
databases. Nodes are components of networks (generally, where two links intersect) and
, For the modal networks, each database may consist of one or more files that have link, node and geometry
information.
DATA BASES AND COMPUTER SUPPORT FOR RISK ASSESSMENT 247
are not necessarily terminals. Similarly, tenninals are probably not located at a given
node although they will be referenced to the nearest node on the appropriate network.
Note that the locational attributes will probably consist of, or include, the National
Spatial Data Infrastructure (NSDI) core attributes, which have not yet been formalized.
Transportation attributes consist of information used for planning and analysis.
These attributes will probably be maintained in several databases and will be identified
from specific analysis requirements and by the aVailability and quality, of the desired
data.
U.S. Modal Networks. Of the three principal modal networks, the highway network
data base consists of approximately 400,000 miles of highways, or 10% of the total
roads in the United States. The rail network contains the entire mainline and secondary
rail systems while the waterway network includes all navigable waters.
Maintenance. The information within each of the databases will be maintained by the
appropriate agency. BTS will maintain selected transportation attributes that are not
maintained by another agency and will support other efforts as requested. For archival
purpose, BTS will maintain archival copies of all information used in the BTS spatial
248 CHAPTER 8
data framework, including the National Transportation Atlas, and for ~y analysis
performed by or for BTS.
Atlas Databases. The BTS will provide annual compact discs of the National
Transportation Atlas databases, initially at two scales: 1:100,000 for local and regional
purposes, and 1:2,000,000 for national evaluations. The discs will contain all associated
databases in ASCII format, the corresponding documentation, and display and
extraction software beginning with Version l.0 in 1995.
Conclusion
Spatial data and GIS technology are as important as traditional statistics to BTS.
Geographic and temporal variations must be understood to understand transportation
issues. In support of this belief, BTS has committed to compiling, developing and
maintaining a collection of databases that represent the current transportation system
along with the necessary user aides to make this data available to the general public.
This collection is generally referred to as the National Transportation Atlas databases
and will be available to the public during the flI'St quarter of Calendar Year 1995. BTS's
goal is to provide data and analysis tools, but not turn-key solutions to transportation
issues.
Different computer support systems for risk assessment and management of dangerous
goods transportation problems are presented below. Software for risk assessment, and
transportation risk in particular, is usually complex, requires large amount of input data
of various kinds (data bases) and a considerable amount of effort in both the problem
formulation and the interpretation of the results. Risk assessment often concentrates on
human fatality risk, and sometimes on property damage. There is, however, an in-
creasing interest for including environmental consequences into a truly multicriteria risk
assessment.
The software/codes mentioned below were developed and are sold by Technica 12 in
London:
WHAZAN-ll
It is a chemical hazard analysis software package designed to run on an IBM PC. It
helps in assessing the toxic and flammable consequences posed by the release of
hazardous industrial substances using simple analytical methods. Programs predict gas,
liquid or two-phase releases from a leak; thermal radiation from burning pools, jets, or
250 CHAPTER 8
fireballs, explosion blast effects; pool spreading and evaporation; and plume rise and
dispersion. A physical and thermodynamic properties data base is included in the
package for 20 commonly occurring hazardous chemicals. Users may add extra sub-
stances to this library, if required. WHAZAN-ll contains also a sophisticated dense
gas/passive dispersion program, which is particularly useful in simulating the dispersion
behavior of pressurized or refrigerated gases, such as ammonia, chlorine, propane, etc.
The mathematical models included are:
- Liquid Outflow
- Gas Outflow
- Two-Phase Outflow
Spreading & Evaporating Liquid Pool
- Jet Dispersion
- Adiabatic Expansion
- Plume Rise
- Dense Cloud / Gaussian Dispersion
- Gaussian Puff Model
- Passive Plume Dispersion (uses a Gaussian dispersion model with a point source
which can be at an elevated position)
- Burning liquid pool
- Burning fluidjet
- Fireball and BLEVE
- Exploding vapor cloud
- lndoorconcentrationsldoses
SAFEI'I
It is an internationally recognized package for quantitative risk analysis. It automates
the risk assessment of chemical and petrochemical facilities, where toxic and flammable
materials are manufactured, stored and transported. SAFETI combines a complete
library of rigorous mathematical models which, either singly or in combination are used
to derive the individual and societal risk associated with the release of hazardous
chemicals. SAFETI automatically analyses the complex sequences from accident
scenarios, taking into account local population and weather conditions. The output is in
the form of individual risk (contours) and societal risk (F-N) curves.
TECKJEI'
It is a program for modeling continuous jet releases. It assesses the effects of toxic and
flammable clouds produced by pressurized. releases from a vent stack or flare, if the
cloud ignites TECJET shows the extent of the effects. Concerning toxic clouds,
TECJET shows the effect on local population centers.
DATA BASES AND COMPUlER SUPPORT FOR RISK ASSESSMENT 251
CARA
It contains a set of modules for conducting various forms of reliability analysis, i.e., the
module FMECA is used for failure mode effect, and criticality analysis, and the module
CAFTAN for fault tree analysis.
PHAST
It enables the engineer to identify situations which have the potential for causing
unacceptable consequences to the population and the environment PHAST is derived
from SAFETI:
LPG-l model
A model LPG-l has been described by Van Erde et al. 13, which can assess the risks
from transportation accidents related to LPG. The model's first module contains a
representation of the spill mechanism following a transportation accident It determines
the amount of vapor flash-off and the size of any remaining pool based on the
atmospheric conditions at the time of the accident, the amount of pre-heating of the
container, and the thermodynamic properties of the material shipped. Subsequently,
simplified fireball, vapor cloud explosion and BLEVE models are used to compute the
extent to which pre-specified levels of shock-wave and heat radiation damage will be
reached and the radius within which direct flame exposure will be experienced. During
each phase of the analysis, the model considers relevant site-specific mitigating factors,
such as the wind speed, air stability, combustion characteristics and the probability of
encountering ignition sources.
RISKMOD
The computer model RISKMOD was developed to provide assistance in the evaluation
of policies for the movement of dangerous goods, by estimating objective risk. 20 The
model represents individual vehicle shipments of dangerous goods on the truck and rail
networks for which the associated risks are estimated in a series of steps:
• The first step involves an accident rate prediction, followed by a spill rate prediction,
given an accident.
• In subsequent steps, the damages are evaluated of an accident alone, and of an
accident followed by a spill of goods.
• The final tables provide link by link risk estimates and a summary of the total risk for
the entire route.
A novel feature of the RISKMOD model is the valuation of consequences, which
provides a common denominator in terms of expressing risk, so that all risks may be
summed on a route.
offers the user the choice between railways and roads. Risk analysis for each route
segment is performed using the SAFETY package, which includes as output F-N curves
and risk grids. Risk contours can also be generated
In this section we present some selected papers showing GIS applications related to
mapping and spatial analysis.
In the past, many State agencies have focused information resources efforts on
administrative processes internal to their organizations. In the USA (e.g.,Texas) there is
a strong interest in developing a business plan "to prioritize and schedule the develop-
ment of public funded programs for the acquisition of geographic information of
common interest and to seek partners in the public sector for data development". It is
believed that an open systems environment is necessary for the different U.S. States to
be able to fully utilize and coordinate the various data bases associated with such
project. Each U.S. agency, while continuing to be responsible and accountable for its
primary mission, must also support the open systems concept in such a manner that
allows for the easy accessibility and use of GIS data layers by all other governmental
bodies.
In Switzerland, a Federal Confederation, where the different Cantons have important
rights to decide an govern themselves, there are many problems concerning the
coordination of such enterprise. It should be the role of the Swiss Federal Government
to create an advisory commission to coordinate and promote GIS activities, and to
develop what we could call a "State Strategic Plan for Information Resources
Management". Switzerland is now in the proce~s of preparing digitized maps at the
cantonal level to suit the needs of public uti.lities and environmental protection.
Integration and accessibility of the different data bases is not yet as advanced as in the
USA.
A strong GIS infrastructure will enable the creation of a map that encompasses all of
the elements needed to assist agencies when performing duties that: 1) enhance the
economic well-being of the state, 2) protect the environment, 3) conserve and manage
the vast natural resources, 4) provide new human resource opportunities through
planning, and 5) protect the lives and property of the citizens. The greatest lasting
benefit of GIS will come from assuring the availability of high-quality data to all state
workers.
The Texas Department of Transportation (TxDOT) recently fmished digitizing its
County/City map series and has made it available for distribution through the Texas
Natural Resources Information System (TNRIS). The County map series is the first
digital statewide data layer made available as a result of the coordination efforts of the
Geographic Information System Planning Council (GISPC).
DATA BASES AND COMPUTER SUPPORT FOR RISK ASSESSMENT 255
The following papers give information on mapping and spatial analysis using GIS
technology (Figure 8.1).
1) Stockwell, J.R.; Sorensen, 1.W.; Eckert, J.W.; Carreras, E.M., The U.S. EPA
geographic infonnation system for mapping environmental releases of toxic chemical
release inventory (fRl) chemicals, Risk Analysis 13, (1993), pp. 155-164.
2) Von Braun, M., The Use of GIS in Assessing Exposure and Remedial Alternatives at
Superfund Sites, in M.F. Goodchild, B.O. Parks, and L. T. Steyaert (eds.), Environmental
Modelling with GIS, Oxford University Press, Oxford, (1993), pp. 339-347.
3) Wadge, G.; Wislocki, A.; Pearson, E.J.; Whittow, J.B., Mapping natural hazards with
spatial modelling systems, in P.M. Mather (ed.), Geographical Infonnation Handling -
Research and Applications, John Wiley & Sons, Ltd., (1993), pp. 239-250.
4) Wadge, G.; Wislocki, A.; Pearson, E.J., Spatial analysis in GIS for natural hazard
assessment, in Goodchild, B.O. Parks and L.T. Steyaert (eds.), Environmental Modelling
with GIS, Oxford University Press, Oxford, (1993), pp. 332-338.
GIS tools are beeing increasingly used in applications concerning transportation, for
representing ecosystems and other environmental resources, and display the range
affected by the consequences resulting from an incident scenario (e.g., using cercles to
represent the surface area of a possible spill, or plotting corridors along a given road
track). The papers listed below give an idea of such applications:
1) Barkan, C.P.L.; Glickman, T.S.; Treichel, T.T., Tank car spills into aquatic eco-
systems, paper presented at the 1993 TRB Hazmat Transportation Conference,
Albuquerque, New Mexico, USA, (1993).
2) Brainard, 1.S.; Lovett, A.A.; Parfitt, J.P., Using GIS for the risk assessment of aqueous
waste transport in Southern England, paper presented at the Society for Risk Analysis
annual meeting, San Diego, California, (December 1992).
3) Brainard, J.S.; Lovett, A.A.; Parfitt, J.P., GIS as a risk assessment tool for aqueous
waste transport, in EGIS'93: Proceedings of the Fourth European Conference and
Exhibition on Geographical Infonnation Systems, Utrecht, EGIS Foundation, (1993), pp.
391-400.
4) Freckmann, P., Route calculation for dangerous goods transport with a graphical
infonnation system, in EGIS '93: Proceedings of the Fourth European Conference and
Exhibition on Geographical Infonnation Systems, Utrecht, EGIS Foundation, (1993), pp.
1132-1138.
5) Lepofsky, M.; Abkowitz, M.; Cheng, P., Transportation hazard analysis in integrated
GIS environment, Journal of Transportation Engineering, 119, No.2 (1993), pp. 239-254.
6) List, G.F.; Turnquist, M.A., Routing and emergency response team siting for high-
level radioactive waste shipments, Paper presented at the 1993 TRB Hazmat
Transportation Conference and subsequently for the special issue on Emergency
Management Engineering of the IEEE Transactions on Engineering Management, March
1994. (Available only from authors, as of 10 May, 1995).
7) Padgett, D.A., Assessing the safety of transportation routes for hazardous materials,
Geo Info Systems 2 (1992), pp. 46-48.
8) Parfitt, J.P., et al., The Risk Management of Hazardous Wastes, Their Transport and
Disposal. Report to the British Oxygen Chemical Foundation, University of East Anglia,
(1993).
9) Patel, M.H.; Horowitz, A.J., Optimal routing of hazardous materials considering risk
of spill, Transportation Research, A 28, (1994), pp. 119-132.
Somer papers presented in this section demonstrate interesting methods for modelling
transportation risk, optimizing routing of dangerous goods, or evaluating the economic
of alternative routing strategies:
1) Abkowitz, M.; Eiger A.; Srinivasan, S., Estimating the release rates and costs of trans-
porting hazardous waste, Transportation Research Redord 977, Transportation Research
Board, NRC, Washington, D.C., (1984), pp. 22-30.
2) Current, J.R.; ReVelle, C.S.; Cohon, J.L., An interactive approach to identify the best
compromise solution for two objective shortest path problems, Computers and Operations
Research 26 (1990), pp. 187-198.
3) Glickman, Th.S.; Sontag, M.A., The tradeoffs associated with rerouting highway
shipments of hazardous materials to minimize risk", Risk Analysis 15(1) (1995).
4) Gopalan, R.; Kolluri, K.S.; Batta, R.; Karwan, M.H., Modeling equity of risk in the
transportation of hazardous materials, Operations Research 38(6) (1990), pp. 961-973.
5) List, G.; Mirchandani, P., An integrated network/planar multiobjective model for
routing and siting for hazardous materials and wastes, Transportation Science 25(2)
(1991), pp. 146-156.
6) ReVelle, C.; Cohon, 1.; Shorbrys, D., Simultaneous siting and routing in the disposal
of hazardous wastes, Transportation Science 25 (1991), pp. 138-145.
7) Saccomanno, F.F.; Chan, A.Y.-W., Economic Evaluation of Routing Strategies for
Hazardous Road Shipments, Transportation Research Record 1020, Transportation
Research Board, NRC, Washington, D.C. (1985), pp. 12-18.
Figure 8.3. Refences concerning modeling and optimization methods for transportation
DATA BASES AND COMPUTER SUPPORT FOR RISK ASSESSMENT 257
Figure 8.3. Refences concerning modeling and optimization methods for transportation (continued).
Summary
This chapter is dedicated to the presentation of data bases and computer support
systems for risk assessment. A number of relevant and specialized databases are
identified and summarized. Details are given on the US transportation statistics and
information on commodity flow. Spatial data and Geographic Information Systems
(GIS) are integrated within the computational framework of transportation risk of
dangerous materials. Computer software is intensively used for the calculation of risk
resulting from a given transportation system. Relevant software was identified and
summarized. Risk and cost analysis models are discussed in the context of risk
assessment tools for the decision making process.
258 CHAPTER 8
References (Chapter 8)
Demonstration Prototype System. (Final Report to EURATOM CEClJCR, Ispra. Italy). Reprinted as CP-
86-10, (1986). International Institute for Applied Analysis, A-2361 Laxenburg, Austria.
23 Kleindorfer, P.R.; Vetshera, R., Risk-Cost Analysis Model for the Transportation of Hazardous Substances,
Final Report VoL I. October 1985, Ludwig Boltzmann Institute, Vienna.
DATA BASES AND COMPUTER SUPPORT FOR RISK ASSESSMENT 259
24 Majchrzak. 1., DISCRET: An Interactive Package for Multicriteria Decision Making Problems, Report,
June 1985, System Research Institute, Warsaw, Poland.
2S Fedra, K.; Lassarre, S.; Weigkricht, E., Computer-Assisted Risk Asssessment of Dangerous Goods Trans-
portation Case Study for Haute Normandie. Final report submitted to the Institut National de Recherche
sur les Transports et leur Securite. International Institute for Applied System analysis. A-2361
Laxenburg, Austria.
26 TECHNICA: The SAFETI Package, Computer-based System for Risk Analysis of Process Plant. Vol. I-IV
and Appendices I-IV, (1984), Technica Ltd., Tavistock Sq., London.
CHAPTER 9
One objective of hazard assessment is to quantify the hazard and the risk of its
occurrence. Another is to reduce the risk.
• The probability of the event can be made low by the application of measures
which can be shown by a simple risk analysis to be of equivalent safety to the
normal standard or code of practice (equivalent risk).
• The probability of the event must be assessed quantitatively by use of (a) a Risk
Analysis, or (b) a Fault Tree (and eventually an Event Tree), and (c) Cause-
consequence diagram.
In some cases events with rather low probability may be neglected, while in others
they still make a significant contribution to the risk and must be taken into account.
A risk analysis may also be carried out to determine safety relative to an absolute
risk criterion such as FAR (Fatal Accident Rate). Whereas afault-tree is constructed by
defining a top event and then defining the cause events and the logical relations
between these cause events (top down approach). An event-tree begins with a particular
initial event and works from the bottom up.
• On the basis of HAZOP studies prepare an overall survey considering major
hazards such as large explosions or toxic release, and the events which can give
rise to them.
• Represent the situation revealed by this overall survey and its possible causes in
the form of a fault tree. The individual branches of this overall fault tree are then
considered and their probabilities assessed.
• On the basis of the hazard assessment and in relation to the risk criteria take
measures for the control of the hazard and for the reduction of the hazard risk.
One problem of risk assessment is that there is no method of ensuring that all
eve.1tualities have been taken into consideration. AnotLer problem is that the data
required for the assessment are often inadequate or non-exi!:tent.
(9.1)
INTEGRATED RISK / SAFETY MANAGEMENT AT REGIONAL LEVEL 263
where pO is the peak overpressure; r the distance; and W the mass of explosive; n =
1.6 (for overpressures of 1 - 10 psi), respectively, 2.3 (for overpressures of 10 - 100
psi).
b) For toxic release
(9.2)
where Q is the rate of release; r the distance; and C the toxic gas concentration. The
value of n is about 1.76.
In many situations it is necessary to relate the intensity of some phenomenon such
as heat radiation from a fire, overpressure from an explosion or toxic gas concentration
from a toxic vapor release to the degree of injury or damage which can result from it. A
method currently used is Probit Analysis. The probit (or probability unit), Y, is related
to the probability, P, by the equation:
1
P=---J Y-5 (u )
exp - - du
2
(9.3)
(21t )1/
2 - 2
The probit is a random variable with a mean 5 and a variance 1. The probability
range (0 - 1) is usually replaced by a percentage (range 0 - 100). A general form of the
probit is: Y = kJ + k2 In V, where the probit, Y, is a measure of the percentage of the
vulnerable resource which sustains injury/damage, and the variable, V, is a measure of
the intensity (overpressure, etc.). Probit equations have been developed for all kinds of
hazards and for different toxic releases.
impacts to be considered
material balance consideration in terms of risk
per net production (e.g., energy) output
acceptable threshold criteria.
• At a regional level, the setting of boundaries will greatly benefit from structured
public scoping/participation process, so that all concemed::ould be involved in
boundary setting. This process would greatly enhance the credibility and accep-
tance of the study process.
Usually in a region where complex risk assessment studies are made, one has to take in
consideration that a large number of compounds are released into the atmosphere and
water and disposed on land with associated environmental impacts (risks). These
impacts may be immediate or long term. Implicit "environmental impacts" are often a
direct or indirect "inference" of "health impacts". TABLE 9.1 and TABLE 9.2 show a
generalized environmental transfer model outlining the various essential components of
health and environmental risk estimations in large industrial complexes/regions.
Transportation of dangerous goods systems should be included and/or integrated into
such a comprehensive analysis.
TABLE 9.1. Categories of risks usually adopted to assess and compare the health impacts of different
pollutant sources
Health Risk
Source People at Risk Exposure Effects
Routine or accidents Workers and the Short, or medium and Fatal and non-fatal
public long term Immediate/delayed
respectively
Long-term/delayed
TABLE 9.2. Categories of risks usually adopted to assess and compare the
environmental impacts of different pollutant sources
Environmental Risk
Source Effects
Duration Extent
Routine or Short or medium and Local, regional and
Accidents long term global
A linear pollutant pathway modelS indicating the amount reaching the receptor
(target at risk) as a function of the amount emitted, altered by dilution and removal and
enhanced by environmental accumulation factors is shown in Figure 9.1.
266 CHAPTER 9
As shown, distance (space) and rate of movement (time) are critical parameters to an
environmental impact assessment study. The complex interdependencies between time,
space and feedback mechanisms (the degree of resilience of a given environment to
external factors) are not fully known; very often it is difficult to normalize them on a
common scale for comparison.
In comparative risk assessment of different technologies one has to "compare risks",
which may be different in a subjective way from impacts, effects, emissions, etc. Two
major limitations have to be considered when dealing with the assessment of environ-
mental impacts of different technologies or technological chains, namely:
- the effects are not always susceptible to quantification
- there is no general agreement on what should be quantified.
Methods of relevance for making comparisons of environmental impacts in
integrated regional risk assessment and safety management studies are:
ranked matrix environmental assessments
emission values and ambient quality indices
critical loads and critical levels.
Rate of Effects
Raleof Rate of removal or
altha
emission transport accumula-
receptor
tion
Amount
Amount
reaching
emitted
receptor
over
over
unHtime
unHtime
Figure 9.1. Pollutant pathway model used for assessing environmental impacts from emissions.
N
~
268 CHAPTER 9
The critical loads approach for regional environmental impact assessment requires
the following types of information:
inventories of current emissions and projections of future emissions rates
estimates of the potential emissions reductions
long range transport models
maps of critical loads and target loads
integrated assessment modeling.
Indicators of environmental risks: Qualitative and quantitative indicators could be
used for environmental impact indicators of different technologies which might coexist
in a region (including the transportation of dangerous goods ). Such indicators are: site
specific, society specific, value judgment dependent. Synergistic effects could be taken
into consideration when dealing with comparative environmental impact assessment.
Environmental performance measurements already accepted by international orga-
nizations and practice and which can be used in comparative analysis are:
environmental performance indicators (e.g., river quality, air quality, soil quality
etc.)
- environmental goals (critical loads, sustainability index)
environmental emissions (SOx, NOx, CO2, etc.).
When dealing with comparative risk assessment for activities which take place in large
(regional) areas, including the transportation of dangerous goods, uncertainties occur in
the prediction of scenarios, models and data.
Uncertainties in scenarios
- erroneous probabilities
- factors not considered
- factors screened out.
Uncertainties in models
- imperfect conceptual model
- imperfect mathematical model
- imperfect computer model.
Uncertainty in data
- general vs. site specific data
- measurement errors
- data reduction.
Expert judgment is inherent in the evaluation of uncertainties. Uncertainties must be
delineated and exposed whenever appropriate and attempts made to deal with them.
Uncertainties take many fonns and it is essential that a coherent and clearly visible
approach is adopted, both in the computational process and in the interpretation of
results.
270 CHAPTER 9
Within the complex study process of integrated regional risk assessment and safety
management (including the transportation of dangerous goods), the uncertainties arise
from a number of sources:
Data. All data are subject to sampling errors. Statistical uncertainty requires the
reporting of range or confidence levels.
Models. Estimates of risks through scientifically based models (lack of direct observa-
tions) brings additional uncertainty in a comparative risk assessment. When models are
the only tool to be used for risk assessment, uncertainty analysis of the results should be
exposed and taken into consideration.
It is clear that an increase in any of these six elements leads to higher estimates of
risk levels. There is always some uncertainty about each of these elements in the risk
estimation process. In the current discussion, the terms "extreme certainty" and
INTEGRATED RISK I SAFETY MANAGEMENT AT REGIONAL LEVEL 271
"extreme uncertainty" denote opposite ends of the spectrum of how well a particular
phenomenon is understood. Qualitative modifiers "high", "fair", and "moderate" depict
gradations of this understanding. These gradations, though arbitrary, are appropriate to
the range of uncertainties found in environmental risk assessment.
1) Uncertainties on the probability of a release of a harmful substance:
- Such estimates improves with the quantity of the information gathered on the
harmful substance released into the environment.
3) Dispersion of a harmful substance and uncertainties on the resulting concentrations
of that substance in the environment:
In practical situations estimates of concentrations at particular points in the
dispersion pathway range from fairly to highly uncertain.
- When measurements of concentrations are available, risk analyst can omit
estimation of the probability. of release of a given substance.
- Actual measurements of concentrations of particular substances in the environ-
ment reveal that concentrations are highly variable ambient concentration of
some pollutants are moderately invariable to moderately variable from day to
day in a particular season. These concentrations may vary by a factor of 5 to 10
between warm and cold seasons and among different years. Human exposure to
some chemicals in the workplace may be extremely variable among individual
work situations.
4) Uncertainty on the population exposed to release of a harmful substance:
In the absence of site-specific data, any estimate of the population potentially
exposed to the release of a harmful substance from such sites would be extre-
mely uncertain.
5) Uncertainties on the uptake of harmful substances by humans and other organisms:
- Pharmaco-kinetic models reflect how the physiology of humans differs from that
of test animals with respect to uptake, metabolism, and excretion of particular
chemicals. Pharmaco-kinetic models may modify risk estimates in significant
272 CHAPTER 9
ways, but their structure and the data they contain make these modifications
moderately to extremely uncertain.
6) Uncertainty on the relationship between dose of a harmful substance and adverse
toxicological response:
The toxicity of a particular dose of a substance not only varies across species,
but among individuals of the same species.
Most acute toxicity estimates are moderately uncertain.
Cancer potency estimates are fairly to extremely uncertain.
7) Measurement error:
Measurement error increases the uncertainty associated with each of the above
elements in an environmental and health risk assessment.
Among all the elements of risk assessment, measurement error generally intro-
duces the least uncertainty.
8) Overall uncertainty:
Uncertainty in each of the elements of a quantitative risk assessment combines to
produce an overall uncertainty in the estimate of a particular risk.
Only a few studies have carefully integrated the uncertainty associated with
individual elements of a risk estimation into an overall estimate of a risk and its
overall uncertainty.
Estimates of human cancer risks due to exposure to chemicals have been shown
to be moderately to extremely uncertain.
9) Credibility of Risk Analysis; Uncertainty on Policy Implications:
- Uncertainties in environmental and health risk assessment are present in all
elements of the assessment process. .
There are several approaches to deal with uncertainty:
i) Risk analysts should present the particular level of uncertainty in the risks they
are estimating.
ii) Risk managers should reduce uncertainty when the reductions would be cost-
effective and would affect the choice of risk management option.
iii) When conservative assumptions must be applied, this must be done in the
context of the particular uncertainties of the risk being addressed.
Reducing the uncertainty in anyone or more of the elements of a risk assessment
will reduce the overall uncertainty of a risk estimate; not all such reductions are equally
cost-effective. Risk managers can also use the economic concept of the efficient frontier to
determine which reductions in uncertainty are likely to reduce overall uncertainty the most
and at the lowest cost:
with regards to making conservative assumptions in the context of uncertainty,
risk mangers should consider when and how to apply these assumptions
INTEGRATED RISK I SAFETY MANAGEMENT AT REGIONAL LEVEL 273
TABLE 9.3 presents some verbal description pertaining to the range of uncertainty.
TABLE 9.3. Range of uncertainty and variability in environmental risk assessments
Acceptability. The assumptions for risk estimation should be discussed when necessary
in a forum (e.g., steering committee).
Lack of a single integrated risk indicator; the elements of risks from various technolo-
gical transportation systems can be broadly categorized as indicated in TABLE 9.3. In
reviewing risk assessment and management results for their use in area safety planning,
one should take in consideration the followings:
• The various elements and dimensions of environmental and health risks cannot be
integrated into one overall indicator of total risk. No comparison on the basis of a single
indicator is possible. The comparative risk assessment process must specify on which
basis (indicator) the comparison is being made.
• It is necessary to expose all the dimensions and elements of health and environmental
risks in thl! comparative process. Differences between transportation systems, regions
and societies make the development of one overall indicator meaningless. There is no
"global" risk value. The results should not be transferred from one study to another
without appropriate investigations of differences between technologies, regions and
countries.
• One approach to develop an overall indicator is the estimation of "extemal costs" of
impacts in terms of monetary value.
• The integrated approach to health and environmental risk means all risks should be
identified, assessed and considered in the comparative risk and management process.
276 CHAPTER 9
External Costs. They are defined to be the costs which fallon one group of people
due to the social, or economic activities of another group, and where the latter group
does not take these costs into account. There are often difficulties in defining to what
extent damage to health or the environment is an external cost.
In order to be comprehensive, consistent and transparent it is necessary to consider
the environmental impacts and their monetarized values over all space and time (the life
cycle analysis). All part of the analysis are subject to uncertainty, some of which is very
considerable. In many cases it is not even possible to describe the uncertainty in terms
of distribution of values. The valuation of uncertain events (risk evaluation) poses
INTEGRATED RISK I SAFETY MANAGEMENT AT REGIONAL LEVEL 277
Public health
- PM IO mortality - PM IO respiratory hospital admission
- S~ mortality - S~ cough episodes for children
- PM IO bronchitis in children - S~ adult chest discomfort days
- PM IO symptoms days - SO, eye irritation
- PM IO asthma attack - NO, eye irritation
- PM 10 restricted activity day - Transport accidents
- PM IO emergency room visits - Construction accidents
Occupational health
- Mining (respiratory diseases; mortality - Transport (death, major/minor injuries)
and morbidity)
- Mining (death, major/minor injuries) - On road accidents (death. major/minor
injuries).
Impact Categories
A list with guiding priority impact categories is given next:
atmospheric poUutionlhuman healthlmaterialslforests/cropslfreshwater ftsheries
accidents (continuous emissions) occupational and public
land use/natural ecosystems
278 CHAPTER 9
global warming
emissions to water/drinking water/ecosystem
noiselhuman health/amenity.
The assessment and hence the comparative risk assessment (CRA) of different transpor-
tation technologies or appropriate technological systems as regard major accidents,
focused almost entirely on estimating acute fatalities to people from historical records
or using techniques of Probabilistic Safety Assessment (PSA) as a predictive tool.
Methodologies for the estimations of the late health effects to people and of the
environmental impacts from major accidents are limited, or in need of significant
development and applications. A defmition for severe accidents should encompass all
the elements of health, environmental risks and damage to plant, equipment, buildings
and should be expressed in terms of both the potential as well as the actual damage and
risk.
The comparison cannot be made on the basis of the consequence of such accidents in
isolation. The likelihood (or probability) of occurrence should also be taken into
account. Hence, estimation of the frequency of such accidents is relevant. Such
estimations necessitate reliable information on the past records of such accidents and
their effects and/or the application of probabilistic methods that predict the likelihood of
their future occurrence.
It is difficult to assess and compare the frequency and the health and environmental
damages caused by severe accidents because such data are not systematically collected
by a single national agency.
There are no data concerning in particular the delayed effects on health from severe
accidents for non-nuclear energy systems. All health effects in such cases are reported
in terms of immediate fatalities, with immediate injuries reported in a few cases. This
makes complete comparison difficult, since the total impact may be underestimated.
The ultimate long term environmental effects, particularly from severe accidents, are
difficult to establish; it may be difficult to establish whether the effect is irreversible or
whether a recoverable effect is possible.
INTEGRATED RISK I SAFETY MANAGEMENT AT REGIONAL LEVEL 279
It is already acknowledged that the process of initiating, promoting risk analysis and
implementing safety management studies for regional areas involves complex decisions
as well as the participation of a large number of actors (e.g., public, experts, systems
analysts, environmentalists, safety engineers, administrators, politicians etc.). The
process of integrating various aspects of risk such as environment, health, hazardous
installations, transportation systems, safety culture, management, involves decision
aiding techniques which are close to the management science field known as decision
analysis. 6
There are, in general, positive and negative aspects associated with decisions.
Indeed, many decisions are made intuitively by experts and do not use structured
processes or techniques. For many decision problems related to risk assessment and
safety management in large and complex industrial areas, the solutions and then
advantages and disadvantages may not be immediately apparent because of the
complexity of the issues involved. There is a need for systematic processes to be
followed that help structure thinking and analysis, and allow different viewpoints to be
taken into consideration. Structuring helps avoid inappropriate ad hoc decisions and
allows the process of reaching a decision to be more open and the decision itself to be
more readily defensible (decisions which are taken today very often have long term
effects). In the end, the use of various decision aiding techniques and the overall
process and technology of decision analysis allows ·the integration of various risk at
regional and area level. The integration of various risks into the decision making
280 CHAPTER 9
process is the appropriate mechanism which allows displaying various risks and
choosing the most appropriate safety improvement strategy.
There are many inputs, influences and constraints that a decision maker will
consider when deciding whicb actions to initiate regarding risk reduction or safety
management to a particular plant or for the region under analysis. Decision aiding
techniques (DAT) are tools for decision makers; they are decision aiding techniques,
but not decision making techniques. In the content of regional safety management such
techniques could i) assist structuring the complexity of issues derived from hazard
identification and prioritization, ii) assist in improving safety of individual installations,
iii) siting of new installations to comply with numerous criteria and performance risk
indicators, etc. There are complex decisions that need to be made in the field of regional
safety management which may involve conflicting positions. A large number of tools
are available to assist in solving and structuring decisions of such complexity.
The main purpose of this section is:
- to emphasize and strengthen the use of DAT for integrated regional risk
assessment and safety management, including the transportation of dangerous
goods.
to provide guidance as to where such processes mayor may not be appropriate
(to avoid the application of sophisticated processes to trivial problems).
- to recognize that DATs are tools, not panaceas.
The main stages of DAT in relation to the integrative process of various risks which
could coexist in a given region, are:
.~: Definition and description of the problem.
.~. Consideration and defuntion of appropriate quality ~surance requirements.
• ~: Formalizing the descriptive model of the problem.
.~: Obtaining the necessary information for modeling.
• Step 5: Analyze in order to determine the set of alternatives and criteria.
• Step 6: Ensure the selection of the proper method to make the decision regarding the
proper integration of various risks in the region and their minimization
• Step 7: Establish a clear record of the process and any decisions taken as a result of
the integration process of various types of risk and appropriate safety
measurements.
To make a decision means to select a way of action out of a set of possible alterna-
tives. The decision process is complex and sometimes iterative; the set of alternatives or
criteria may vary from an interaction to iteration.
Decision making involves three major elements:
a) alternatives among which the ''best'' one will be chosen
b) criteria for judgment
INTEGRATED RISK I SAFETY MANAGEMENT AT REGIONAL LEVEL 281
Before carrying out any decision aiding process for integrated risk assessment and
safety management at regional level, it is essential to identify all of the available options
and relevant factors that would influence the outcome. "Brainstorming" can present a
constructive approach. The use of multi-disciplinary scientific and engineering
expertise groups should be considered for dealing with the more complex assessment
problems, particularly those involving the need to choose between aVailability of
alternative or competing action pathways.
Generally, one option is devoted to a given objective function (e.g., reduction of the
probability of occurrence of an undesired event, reduction of various releases, etc.). One
can thus structure a brainstorming session as follows:
What are the involved functions?
- Is each function necessary?
- It is possible to ensure a given function by different ways?
- For each way, is it possible to ensure the function differently?
The selection of the relevant factors can be aided by the application of analytical
hierarchy process or value trees which enable the complexity of a given problem to be
broken down into smaller constituent parts. As an example, TABLE 9.1 and TABLE
9.2 illustrate the criteria hierarchy for the overall health risks. As shown, distance
(space) and rate of movement (time) are critical parameters to an environmental impact
assessment study. The complex interdependencies between time, space and feedback
mechanisms (the degree of resilience of a given environment to external factors) are not
fully known; very often it is difficult to normalize them on a common scale for
comparison.
In comparative risk assessment of different technologies one has to "compare risks",
which may be different in a subjective way from impacts, effects, emissions, etc. Two
major limitations have to be considered when dealing with the assessment of
environmental impacts of different technologies or technological chains, namely:
INTEGRATED RISK I SAFETY MANAGEMENT AT REGIONAL LEVEL 283
A number of decision-aiding techniques are currently available. For the purpose of this
book, the application of Cost-Effectiveness Analysis (CEA), Cost-Benefit Analysis
i.CBA), Multi-Attribute Utility Technique (MAUT), and Multi-Criteria Outranking
Technique (MCOT) are considered in further details.
Any decision aiding procedure involves the use of data, models, techniques, value
judgments which contain uncertainties of various types such as:
Uncertainties associated with an imperfect knowledge of the performances of the
options under different circumstances (scenarios) or of the parameters and data
used in the assessments.
- Uncertainties associated with the use of models.
- Uncertainties due to imperfect knowledge (e.g., about the future of each option).
Intrinsic uncertainties resulting from the statistical treatment of the variable (this
is very important when assessing the expected outcomes from low probability
events).
- Value judgments.
The sensitivity analysis highlights the aspects which have the greatest influence on
the results, where it may be desirable to attempt to reduce the uncertainties, if possible.
Types of sensitivity analysis involve the following process:
- modification of a given parameter.
assignment of probability distributions to the important variables (probability
encoding).
INTEGRATED RISK I SAFETY MANAGEMENT AT REGIONAL LEVEL 285
Aspects to consider when presenting the results of integrated decision for risk
assessment and safety management at regional level are as follows:
List of the assumptions, hypothesis and initial conditions considered within the
case study
Comment on the weighting factors
Comment and reference on the various models and techniques used as well as
their relevance and integration within the overall study
- Indication of the ~ uncertainties, quality of data
Presentation of the sensitivity analysis results.
From these different information the final decision will be made by the decision
maker or the group involved in preparing recommendations and policy strategy at area
level.
Artificial intelligence models have been lately implemented for computer aided design
for risk assessment and management, as well as for expert systems development and
implementation. A few comments are worth considering:
a) Expert systems have been developed for assisting the probabilistic safety
analysis for complex nuclear and other industrial facilities and processes; they
are relevant in integrating various aspects concerning risk in a given region.
b) Decision support systems by means of expert systems, operation research, multi-
criteria decision models with objectives in conflict are already in use in making
decisions concerning risk reduction and cost optimization for marginal safety
measures implementation.
c) Connectionist expert systems (neural networks)7 which exhibit characteristics
and capabilities not provided by any other technology; this a complementary
technique designed to assist solving ill-designed risk assessment and manage-
ment problems.
d) Integrated knowledge based decision support systems by using systems analysis
techniques and procedures and by integrating in an efficient way expert systems
and neural network technology into an advanced methodology of systems
analysis and information technology.
e) Influence diagrams which enables capturing numerous elements which could
lead to an unsafe operational environment
286 CHAPTER 9
Decision Conferencing
Decision Conferencing is an efficient tool for integrated regional health and environ-
mental risk assessment and safety management. Decision Conferencing is a relatively
new decision aiding tool, which has been developed from the practice and need of
integrating various aspects of risk, which has grown up in a number of organizations or
projects, when complex situations and difficult issues arise. In a recent report
("International Chernobyl Project - input from the Commission of the European
Communities to the Evaluation of the Relocation Policy adopted by the former Soviet
Unions) it is highlighted the followings:
"Rather than discuss the matter amid the hurly-burly of their day to day
activities, the decision makers take time to go away from their regular working
place in order to concentrate fully and solely on the issues before them.
Sometimes straightforward discussion can lead to a clear decision and a view of
a way forward, but the complexity of the issues and the uncertainties involved
may be too great for simple discussion to resolve. In such cases, it has become
the practice in some organizations for the team responsible for the decision to
meet together for two days or more away from their nonnal working
environment to discuss and explorer the issues".
Decision conferencing uses the services of a moderator, who is skilled in the process
of group decision. The moderator is seldom a person with experience in the context of
the issues at hand or have expertise in the discussion. The moderator has a very definite
role, namely to smooth the team's work:, to help the process and make the team
productive and more creative. Within a decision conference event, the content of the
discussion comes from the team themselves. A decision conference is a two-day event
in which all the "owners" of the problem (e.g., integrating health and environmental
risks in a given region and making appropriate decisions for safety improvements)
gather together to agree upon a strategy. The decision makers, or advisors (the problem
owners) are supported by a moderator and a decision analyst. Two important comments
are made:
i) the moderator has the role to lead the conference (he acts as a disinterested
chairperson guiding the discussion forward constructively, building and interpreting
decision models): this would help the decision makers or their advisors to appreciate
various facets of the problem before them.
ii) the analyst deals with the details of the model building, by using information
technology (models and computers).
The technology of decision conferencing brings together knowledge, skills and
techniques arising from the following fields:
• Decision analysis which provides a structure and language (e.g., multi-attribute value
theory) in which decision makers can think and talk about their problem.
• Information technology assists in "real time" the decision making process; the results
of decision analysis can be presented to the decision makers in simple, easily
comprehended ways.
INTEGRAlED RISK I SAFETY MANAGEMENT AT REGIONAL LEVEL 287
• Group dynamics an awareness of which helps the moderator and analyst ensure that
the group of decision makers or advisors interact constructively.
• Within a decision conference the team can discover the importance of differences in
opinion between the team members and the sensitivity of their conclusions to these
differences and to those judgments of which they are most unsure. Controversial
discussions are unimportant when discussion is focusing on the critical issues without
being deflected. The moderator is expert in group dynamics and rational decision
theory, as well as a skilled communicator.
In practice, each decision conference is different. It evolves according to the needs
of the team and not according to some fixed agenda. The topic of integrated regional
risk assessment and safety management and its complexity is an ideal topic for decision
conferencing. No model should be taken as definitive: revised models are build and
much sensitivity analysis undertaken until no further insights are obtained and agreed
conclusions are reached. Decision conferences are creative events; they create an
environment where participants create, evaluate, modify and re-evaluate options,
building a strategy which they all support.
In accortiance with the practice in this field: "Surprisingly, the analysis in
decision conferences needs much less hard data than one would, at first, think.
Strategies have to be priced: that is clear. But the cost estimation need only be rough. It
is a broad brush picture that the event seeks to create. Detail can be added at a later
date".
The lack of a single risk indicator to deal with risk assessment, health and environ-
mental impacts at the regional level, should be compensated by the use of this
specialized model - decision conferencing - able to integrate objective and subjective
decision aspects when dealing with complex issues and costly options and decisions.
9.5.1. CAUSEOFLOSSES
When considering the different aspects of regional risk assessment, one must keep in
mind that part of the risk is represented by fixed installations, like factories, chemical
processing industry, warehouses, marshaling yards, loading/unloading stations (for
hazardous products in liquid form), etc., and by the means of transportation related to
hazardous materials (trucks, railroad, ship, barge, pipeline).
A major hazard in modem process plant and in transportation usually materializes
due to loss of containment, resulting in leakage of hazardous liquids or escape of toxic
or flammable gases.
The big three major hazards are fire, explosion and toxic release. Some factors
which determine the scale of the hazard are:
• inventory
288 CHAPTER 9
• energy factor
• time factor
• intensity - distance relations
• exposure factors (reduction of exposure; mitigating effect).
TIle control of hazards is possible only through effective management Some
features which are characteristic of loss prevention are:
• development of techniques for the identification of hazards
• principle of risk criteria and of the quantification of hazards
• development of techniques for the quantification of hazards
• use of the techniques of reliability engineering
• principle of independence in critical assessment and inspections
• critical review and critique of codes and standards or regulations, where these
appear outdated by technological changes.
Planing for emergencies is also an aspect of importance in loss prevention.
Loss prevention is concerned with the avoidance both of personal injury and of eco-
nomic loss. Some costs related to loss prevention arise through failure to take proper
loss prevention measures; others are incurred through uninformed and unnecessarily
expensive measures. The financial viability of new project is often determined by loss
prevention factors.
Cost of Losses
The economic factors affected by accidents include loss of human life, property, insur-
ance, cost of being absent from work, plant shutdown and downtime, loss of marlrets,
etc.
The Concept of "Total Loss Control". The relation between accident and propertl
damage costs is the central concept of "total loss control" (Fletcher and Douglas ,
1971). Here the basic idea is that loss due to personal accidents is only the tip of the
iceberg of the full loss arising from accidents. Indeed accidents can be classified as
disabling injury, minor injury, property damage and no damage accidents. A disabling
injury being one which involves one or more days off work other than the day of
accident. The ratio between the different types of accident is a key concept in loss
control. A ratio for these accidents can be given in the form of an "accident pyramid":
1 (disabling injury) : 100 (minor injury) : 500 (property damage) : 10,000 (no damage).
The modeling of accident process is important and has been studied by Houston
(1971,1977), who proposes the following tools:
• Principal cause model ("proximate cause" used by lawyers and insurers): not
well suited.
• Fault tree model
The following statistics are needed for risk assessment and estimation of the cost of
losses:
- Loss statistics (fire and explosion). Explosion losses are usually included in the
fire statistics.
- Personal accident statistics classified by the type of manufacturing industries
(e.g., injury or deathsl1()8 exposed hours), and the corresponding rates/ratio
should be established as a mean of comparing different types of activities.
Cost of Losses at the National Level. The necessary information can be extracted from:
Statistics for fire loss in factories.
Surveys regarding the national costs of machine repairs and maintenance, both in
terms of direct maintenance costs and of lost production.
The cost of downtime in the process industries.
Information regarding other costs, such as those of design and commissioning,
delays, loss of markets are more difficult to obtain.
Cost of Losses at Factory Level. These can be estimated along the following lines:
• Damage cost may be estimated at the factory level by the method of "total loss
control" 9. Plant unavailabliity or downtime, mayor may not incur serious cost
penalties. There exist methods of estimating both the mean and spread of plant un&vail-
abilities.
• The effect of downtime on plant profitability is another standard topic in process
economics.
• Data on equipment failure are available as generic failure data given in the literature
or as data obtained from the works. Similar considerations apply to equipment repair
times and costs.
• The estimation of the cost due to public reaction is very much a matter of judgment.
This involves also the costs due to legislation changes or changes in the practice of a
particular branch of the industry under public pressure.
• Insurance costs need to be considered both in terms of the rates which are currently
quoted and of those which might be obtained by adoption of further loss prevention
measures.
Loss prevention is in large part the application of probabilistic methods to the
problems of failure in the process industries. Reliability engineering in the process
industries, and in transportation and storage, is especially concerned with mechanical
equipment reliability. Reliability may be defined as "the probability that an item will
290 CHAPTER 9
perform a required function under stated conditions for a stated period of time". We
indicate below some definitions concerning reliability relations:
Probability of unions
Joint and marginal probability
Conditional probability
Independence and conditional independence
Bayes theorem
- The reliability function and hazard rate
- The failure density and failure distribution functions
Relation between basic functions
- . Exponential distribution
Probability and event rate
- Unreliability and failure rate
- Mean life
There is a close link between reliability and qUality control of equipment, welds,
piping and instrumentation. Maintenance and operability controls taking place at
regular interval is an asset for good performance of equipment and is an integral part of
loss prevention.
There is a growing awareness that rational risk management (based on estimates of the
probabilities and consequences of an accident) is only a partial solution to the problem
associated with the risk from the transportation of dangerous goods. Engineering
technology may effectively deal with the problem of reducing real risks; it very often
neglects the perceived or subjective risks. In terms of public response, perceptions are
more instrumental than objective risks. The relationship between ,,real" and perceived
risk is an imperfect one. When saying that risk perception is irrational vis-a-vis
objective risks, one does not imply that there is a lack of logic or pattern to the process.
A number of concerns relating to public perceptions of risk, risk tolerance and risk
acceptance could be considered from the perspective of communication strategies.
When dealing with risk communication programs within a decision making process,
one has to consider the followings:
- Public concerns and priorities
Knowledge and images
- Risk acceptance
Individual and community context.
Public Concerns and Priorities. The first issue is to understand what is bothering
people most about the transport of dangerous goods. One can make the assumption that
a whole range of factors, quite apart from "true" risk, will determine perceptual
INTEGRATED RISK I SAFETY MANAGEMENT AT REGIONAL LEVEL 291
response, and ultimately public acceptance. For the purpose of strategic planning, the
perceived risk is every bit as real and significant as the objective risk.
Public concerns and priorities could be considered from the following prospective:
The relative concern paid to the issue of the transportation of dangerous goods
vs. other threats to health and safety in the investigated area.
Perceptions of the relative risks posed by different modes of transport or types of
goods.
Levels of concern over various consequences of the transport of dangerous
goods.
A number of questions should be considered when dealing with the heading of
public concerns and priorities:
How prevalent and profound is the concern about the transport of dangerous
goods ?
How concerned are people about this issue relative to other threats to health and
safety ?
Is transport by rail more dangerous than by other modes ?
- Are nuclear wastes perceived as more risky than chemical goods ?
- Are people more concerned about risks to health and safety, or the impact of the
transport of dangerous goods on property values and the quality of their living
environment ?
Do these concerns vary systematically according to the background character-
istics of the respondent?
Summary
Integrating various types of risks, and their sources, existing in a region is not an easy
or trivial task. Various types of techniques, methods and tools could be employed in
helping the integration process. One potential I relevant instrument for the integration
process of various types of risks in a region is that of Comparative Risk Assessment.
The chapter investigates the process of setting boundaries for comparative risk assess-
ment of different technologies, the assessment of environmental risks/impacts, com-
parative health risk assessment, uncertainties in comparative risk assessment. Methods
and tools of comparative risk assessment for severe accidents are also discussed.
Recently, Decision Aiding Techniques playa significant role in the integration
process of various types of risks and their management at the regional level. This
chapter reviews such techniques and their relevance to the risk integration process.
Special attention is given to decision analysis and decision conferencing.
294 CHAPfER9
References (Chapter 9)
I Davis, E. I.; Ake, 1. A., Equilibrium thermochemistry computer programs as predictors of energy hazard
potential. In Loss Prevention, 7 (1973), pp. 28.
2 Gordon, S.; McBride, B. I., Computer program for calculation of complex chemical equilibrium
compositions, Report NASA SP-273, (1971).
3 Gheorghe, A.V., Comparative risk assessment of the health and environmental impacts of various energy
systems, In!. I. of Environment and Pollution, 4(3-4), (1994), pp. 329-349.
4 Haddad, S.; Gheorghe, A.V., Issues in comparative risk assessemnt of different energy sources, In!. I. of
Global Energy Issues, 4(3), (1992).
S United Nations Environment Programme, The Environmental Impacts of Production and Use of Energy.
Part IV. The Comparative Assessment of the Environmental Impacts of Energy Sources, Phase L
Comparative data on the emissions, residuals and health hazards of energy sources, Report ERS-I4-85,
UNEP, Nairobi (1985).
6 Gheorghe, A. V., Decision Processes in Dynamic Probabilistic Systems, K1uwer Academic Publishers,
Dordrechl, (1990).
7 Gheorghe, A.V., Connectionist expert systems for analysis of health problems associated with industrial
activity and electricity generation systems, In!. I. Environment and Pollution, 4 (1994), pp. 107-124.
• International Chernobyl Project - Input from the Commission of the European Communities to the
Evaluation of the Relocation Policy Adopted by the Former Soviet Union", Report EUR 14543 EN,
Commission of the European Communities, (1992).
• Fletcher, I.A.; Douglas, H.M., Total loss control, Assoc. Bus. Programmes, London, (1971).
CHAPTER 10
The information presented up to now gives a general overview of the situation on the
transportation of dangerous goods. This includes methodology, statistical data for
probabilistic risk assessment, and other information of interest. There has been different
approaches concerning risk analysis, which were followed by the authorities or
engineering companies in different countries. In this chapter we intend to present a
short summary of selected case studies and miscellaneous information of interest to the
reader, which may show different aspects of the methodologies in use and how a
potential user might implement its own applications. The topics presented below are
intended as examples, and in view of the restricted amount of space, this selection had
to remain very limited, and is in this sense not exclusive.i
In the Vancouver transport risk analysis model, ideas about the evaluation of risk are
organized around two underlying themes: probability and consequences. Comparing
transport routes in order to select the way presenting the least risk, or to focu~; remedial
actions on road segments which are known to be dangerous, is an important task. The
general philosophy followed by the Canadians is highlighted below.
The Government of Canada has sponsored a "transport of dangerous goods study" for the
Vancouver area, and the development of a risk analysis model. The principal issues of this
interesting report are summarized below.
, We wish to apologize in advance to the reader and to any companies, which may feel that a particular
survey should have been included at this place. We also want to thank the Cantonal Authorities and the
private companies, who supplied us with the requested information and allowed us to disclose it in the
present book.
296 CHAPTER 10
results in a rate which is independent of the links road length or the actual number of
vehicles using it. The expected accident rate, ~,is expressed as:
Number of Accidents j
P = . length. (l0.1)
I vehicle j .length j I
The subscript, i, is an index number used to denote a unique value for a specific
section of a roadway.
To estimate the actual number of accidents over an interval of time, it is necessary to
specify the traffic volume.
r .
dangerous commodity may be defined as:
7t • (reach j
'k =value··k · ·modijier·k (10.2)
'J IJ 2 . length .. reach. J
1 J
The three indexes. i, j, and k refer to the road links, hazardous commodities, and
resources, respectively. The value of each resource (valueijk ) along any given link i,
may be derived from the GIS system (Geographical Information System), which can
perform an interpolation based on the information in its database. This interpolation
algorithm accounts for possible mixtures of high and low resource densities along each
link and computes a weighted average of the resource within the corridor defmed by the
reach of each commodity. For any given accident event on a particular link it is
assumed that only the resources within the "circle of effects" are at risk. Therefore in
the equation above, the value of each resource within the commodity specific corridor is
multiplied by the ratio of the area of the circle to the area of the corridor (i.e., the
fraction of the corridor actually at risk from a release of the "standard volume"). Finally
this value is further scaled by the damage effects modifier. The formula for estimating
consequences from a transport accident is given below:
It R2
C iit = V'it ·ITT·Hjt
I J
(10.3)
Term.4. Term B Ter". C
where, the nomenclature and definitions shown in the box were used.
298 CHAP1ER 10
Calculation of Risk
When Eq. 10.1 and 10.2 are multiplied, they estimtJte the amount of a given resource at
risk (in the restrictive sense) on a given link, for a given commodity, when one standard
unit is released. The final evaluation of risk to any particular route is made by
specifying the volume of traffic carrying each commodity along the road link that forms
the route. This volume is defined as the number of "standard volume" units of each
commodity transported along each link per year.
The total risk can then be calculated by adding up the link specific risk for each
commodity of interest, and arriving at a score for each resource, expressed as the total
value of each resource at risk per year.
It is advisable not to express the resources of interest, i.e., people, property, and
environment in a common unit such as the dollar or any other currency value.
Combining the first two equations with traffic volume, the final equation for total
risk to each resource along a route containing 'i' links and l' commodities can be
expressed as:
(10.4)
The reSUlts, which this model produces, are expressed in units which depend upon
the k resources:
Euch of these numbers is typically very small, and so to make them more convenient
to discuss they are given a factor of 1 x 10 as a standard multiplier.
. The risk analysis model describe previously for the Vancouver case study has been
programmed and implemented on an IBM PC/AT micro-computer. The computer
environment enables the user to create and evaluate routes made up of arbitrarily
complex link combinations, carrying arbitrary combinations of commodities along the
great many possible links in the study area. Using an interactive window-oriented
screen, routes and commodities are easily specified.
• . Road movements: map of roads and transport corridors, with associated accident
rates for each section of road, and the volumes of each commodity transported
along each link.
• Population and property: map of concerned areas together wKh population
density and average residential property values for each case.
• Environment sensitivity: broad areas based on watersheW; and land/water cover
with ranked values on a ten point scale accounting foJ' game, bird and fish
habitats at risk.
The GIS is used to provide information about the resources at risk along each link in
the road and rail transport networks. For each item, a map and its associated textual
attribute data must be available. An example of a typical GIS application is shown
below:
For the computation of risks,four products were selected on the basis of their traffic
or their potential hazards: motor spirit (200 trucks/day), LPG (30 truckS/day), ammonia
(2 trucks/day), chlorine (0.2 trucks/day). The route was subdivided into a finite number
of potential accident points (about 20 for each route) to which an accident probability
was attached. A city map was prepared with indication of the population density
(resident and working population) corresponding to selected potentially impacted areas.
The grid was 1 x 1 kIn. In addition meteorological data were gathered and aggregated
according to the wind rose, and three weather conditions were selected.
Key figures for the study concerning the Lyon's area are the release probability per
accident and lethal area knowing the size and nature of a release (6% of total release for
LPG accidents). Lethal area estimation implies hypotheses on release rate, fire ball size,
turbulence, atmospheric dispersion, and on toxicity, or on heat exposure dose response
relationship.
Lethal areas have been computed at the 100% and 10% level according to the 12
weather categories for the four product categories. A few thousand of basic events (e.g..
given physical phenomena following an accident at a given point on the route and under
a given weather category) were taken into consideration. In every case the probability of
the event is known as well as the consequences within each cells, since the number of
deaths is the product of the population density by the fraction of the lethal area within
the cell. Aggregating along the cell provides the total number of deaths due to the basic
event, pooling all the events for a given route provides the risk profile for the route
which can be displayed as a probability - consequences curve or aggregated in an
expected number of deaths. The number of victims among the drivers was also
estimated.
The introduction of weighting factors has been recommended for the US nuclear
industry by the US Nuclear Regulatory Commission (NRC). The weight is the number
CASE STUDIES IN REVIEW AND MISCELLANEOUS INFORMATION 301
of deaths. In Germany, a risk aversion factor for railroad accidents in tunnels was
considered and a weight of ten was assigned to accidents with more then 20 deaths.
In this context, the Swisselectra company has developed a method for connecting
the different parameters together, and has on purpose avoided to define and use accident
scenarios, as this is the case, when accessing the risk of a single event (e.g., spill, gas
release, fife, or explosion).
I I I I
1 I
Magnttudeof Probabiltty of
Potential Losses I Occurence
Damages
L I
I
Risk of a Dangerous
Good's Transport
Among the different factors defining the total transportation risk, there are four main
elements to be considered:
• parameters related to the dangerous goods transported (Physical-chemical
properties, toxicity, energy content, etc.) and the quantities transported.
• parameters describing the environment (soil, water, air). Such aspects like
populationlhabitation density, distance to the next buildings, distance to the next
ground or surface water and size of the concerned protected zone, biotopes,
tunnels, viaducts.
• accident frequency. This describes the number of accidents on a selected road
segment resulting from the road trace, traffic density, type of road, topography,
and other factors influencing the fluidity of the traffic.
• frequency of dangerous goods transports on selected road/rail/waterway
segments.
The first two elements, i.e., product related danger potential and environmental
parameters, describe the magnitude of potential losses as a result of some given
consequences. The last two elements, accident frequency and dangerous goods transport
frequency are a measure for the probability of occurrence of a hazardous event. By
combining these factors we obtain the total risk arising from a dangerous goods
transport (Figure 10.1).
CASE STUDIES IN REVIEW AND MISCELLANEOUS INFORMATION 303
I I
il
Determine the respective danger
potential for each hazardous product
based on the corresponding transport
quantity:
Figure 10.2. Procedure for assessing the potential danger of a dangerous product.
For each selected product the danger potential is assessed, using RID, or TERA lists
(Figure 10.2 and Figure 10.3). The aim is to assess the risk to the different
environmental compartments, PH, PA , and Ps. The potential danger does also include
the dispersion range of a toxic substance.
The resulting danger potential, P, for a given product is obtained as follows:
(10.5)
where
PH : Danger potential concerning air, respectively human beings I animals
PA : Danger potential concerning water
Ps : Danger potential concerning soil, respectively vegetation.
304 CHAPTER 10
I
Classify the substance properties according to danger classes
I
Prepare a finer classification and distribute the substances into further
subclasses
I
Define and adjoin the evaluation criteria
I
Combine these evaluation criteria (ranking factors) with the potential
danger of the substance, in such a way that the resulting figure
represents a measure of action radius and effect range of the hazard
with respect to the different environmental compartments.
Figure 10.3. Procedure for defining the potential danger associated with a substance.
The danger potential for human beings I animals does include the following factors:
(10.6)
where
PI : Inhalation toxicity
PD : Dermatologic toxicity
PE : Potential danger due to the energy content of the product (explosion, fire,
thermal irradiation).
Similar defmitions and algorithms for estimating the danger potential of a substance
relative to water and soil have been defmed in the original Swisselectra report, to which
the reader is kindly referred.
The sensitivity of the environment can only be assessed after the construction
aspects of a road!locality are known, since these may affect the severity of an accident
and influence the retention of spills. A distinction must be made concerning the type of
environmental compartment (air, soil, water, forests) concerned. A matrix procedure is
used to classify and evaluate the respective sensitivities (Figure 10.4).
By combining the potential danger of a substance with environmental data, it is possible
to estimate magnitude of the potential damagesllosses caused to the environment by the
release of toxic products (Figure 10.5). The combination of these elements is defined as
the magnitude of potential losses, or damages. This gives an overview of the critical
points to focus on, along a given road segment concerned with the transport of
dangerous goods.
CASE STUDIES IN REVIEW AND MISCELLANEOUS INFORMATION 305
Compile parameters Apply matrix procedure Compile road/surrounding Apply matrix procedure
describing environmental for classifying lestimating construction aspects for classifying! estimating
characteristics I sens~iv~y environmental sens~ivtty construction aspects
I I I
I
Determine sens~ivtty Determine correcting
of environment factors for Increasing I
decreasing sensitivtty
I
Determine sensitivity of
environment related to:
• Environmental compartments
• Products
• Local conditions
I
Magnitude of Potential Losses I Damages
corresponding to:
• Environmental compartments
• Product transport
• Local conditions
The transport frequency can be directly derived from statistical surveys.and traffic
counting for the given region, respectively road segment The data must be evaluated to
obtain the number of dangerous goods transport per year for selected hazardous
materials.
The Swiss VSS-Norm SN 641 3tOa defines the relative accident rate as:
U·t0 6
Accident rate = _...:'-- (to.7)
D1V; Lj t
where
Uj = Number of accidents on the road segment i
D1Vj = Average daily traffic on the road segment i per kilometer
Lj = Length of the road segment in km
t = Observation period in days.
The average accident rates defined in above mentioned Swiss VSS-Norm are shown
in TABLE 10.1.
TABLE 10.1. Average accidentrales (according to VSS-Norm SN 641 310a)
The accident frequency of road segments, for which no concrete data are available,
can be estimated by taking average accident rates for similar road categories. The
accident rates shall be corrected by applying some correction factors which takes in
account the road particularities (Figure 10.6).
By combining the accident rate (related to a given road segment) with the transport
frequency for dangerous product of concern, it is then possible to estimate the
probability of occurrence of that event (Figure 10.7).
Finally, by combining the magnitude of potential losses/damages with the event
probability, one obtains the transportation risk for the dangerous goods under consid-
eration. Estimation of the risk according to the numerical procedure described here
CASE STUDIES IN REVIEW AND MISCELLANEOUS INFORMATION 307
makes it possible to apply normalized (dimensionless) key values for defining some
characteristic risks and to combine these values (or severity scales) to finally obtain the
total risk.
I
Calculate the corresponding accident
rate for each road segments
I I
I
Probability of occurence
for a given scenario
Figure 10.7. Procedure for assessing the probability of occurrence of a given scenario.
MORIKA is a risk analysis for the Baregg-tunnel l (National highway Nt, Baden,
Switzerland), which has been prepared by "Gruner AG, lngenieuruntemehmung"
(Basle) for the Canton of Argovia The actual status has been investigated and a risk
analysis prepared on the behalf of the Health Department ("Gesundheitsdepartment des
Kantons Aargau, Kantonales Laboratorium").
308 CHAPfER10
Some statistics on road transport of dangerous goods through the Baregg tunnel
(Status June, 18, 1991) are reproduced in TABLE to.2. There were 50 accidents in the
Baregg tunnel for the year 1989, but since this road tunnel exists there has been no
accident involving the transportation of dangerous goods through the Baregg tunnel.
The accidents can be classified as follows:
• Drive up accidents due to heavy traffic
• Accidents resulting from a queue in the tunnel
• Accidents at night due to repairs and maintenance work in the tunnel
• Single accidents during light traffic periods
• No accident involving dangerous goods transports
TABLE 10.2. Statistics concerning the road traffic in the Baregg twmel, Baden. Switzerland
(Status Iune. 18. 1991)
For risk assessment pwpose, incident scenarios were developed to suit the condi-
tions prevailing in this tunnel (see box below):
Effects on People
Degree of Light Medium Heavy Severe
DIIIIlIIpS/
Consequences
groups of consequences, Le., light, medium, heavy, and severe were retained, according
to TABLE 10.3.
Effects on Ecosphere
- Fish dying some dead fishes < 100 kg < 1'000 kg > 1'000 kg
Effects on Infrastructure
Supply Traffic Individual Regional Inter-regional
Interruption damages damages damages
- Effect duration <I hour < 1 day < I week >1 week
- Financial < I Million SFr. < 10 Million SFr. < 100 Million SFr. > 100 Million SFr.
damage
- Effect duration <1 hour < 1 day <1 week > 1 week
CASE STUDIES IN REVIEW AND MISCELLANEOUS INFORMATION 311
A survey of the preventive measures and intervention plans was also prepared, and
the benefit of implementing some new proposed measures for preventing incidents
and/or improving the technical standard and safety of the tunnel were assessed.
New additional security measures will be implemented in the near future, and an
improved coordination of the different instances (police, fire-brigade, hospitals, etc.) is
foreseen.
The Authorities of the Canton Uri have asked the engineering company EB&P (Ernst
Basler&Partner Ltd., Basle, Switzerland) to prepare a risk analysis concerning the
environmental protection status of rivers, ground waters, lakes, and tunnels, what the
transport of dangerous goods by road is concerned. Risk assessment for this area is of
prime importance since the region includes the national highways N2 (from
"Seelisberg" to the "Gotthard tunnel") and N4 ("Axenstrasse"), and tunnels like the
Gotthard tunnel and the Seelisberg tunnel through which an important international
traffic is mowing daily.
We reproduce here the content of EB&P's report, which was summarized and
translated into English to serve as an example of a "Swiss Dangerous Goods Transport
Study".
The statistics concerning the traffic, especially the transport of dangerous goods,
was complemented by direct surveys at some strategic points. The dangerous goods
traffic through the Gotthard tunnel amounts to 7% of the total goods transports through
the tunnel. On a regional basis it represents orJy 5%. The traffic surveys of the strategic
points, for the year 1990 (fABLE 10.4), have shown that the proportion of trucks
carrying dangerous goods can vary between 1 and 3% depending on the road link
concerned and date of the survey. A value of 4% was considered for the risk analysis.
The national highway N2 is an important North-South connection. There are approxi-
mately 600,000 trucks per year transiting on this axis.
TABLE 10.4. Total nwnber of bUcks respectively number of trucks carrying dangerous goods 01\ the
National highways N21 N4 (for the year 1990)
Localities
GGschenen Erstleid Bolzbach Fliielen
(Getthard (Seelisberg
tunnel) tunnel)
VehicleIYear VehicleIYear VehiclelYear VehiclelYear
Statistical Information and Documentation Used for the Transporl Study of the Canton
Uri
• Statistics 1984 on Road Transportation of Goods (Erhebung Uher die Giitertransporte
auf der Strasse 1984. Bundesamt for Statistik).
• Statistics 1989 on Road Transportation of Goods in the transalpine region (Erhebung des
transalpinen Giiterverkehrs 1989. Dienst GVF, Sigmaplan, August 1990).
• Statistics concerning traffic accidents in the canton Uri (Verkehrsunfallstatistik des
Kantons Uri 1986 bis 1989).
• Road traffic accidents in Switzerland 1987 (Strassenverkehrsunfiille in der Schweiz
1?87).
• List of classified liquids dangerous to waters (Liste der klassierten wassergefiihrdenden
Fliissikeiten. BUWAL, 1981).
• Products and Goods Classification. (Klassierung von Stoffen und Waren, Brandverhii-
• tungsdienst BVD).
• Catalog of substances potentially dangerous to water ("Katalog der wassergefiihrdender
Stoffe" des Bundesministerium des Innern der BRD un die schwarze Liste der EG,
1.3.85).
• Guide for road transportation of dangerous goods (Leitfaden for den Transport gefdh-
rlicher Giiter auf der Strasse SDR. Bundesamt /iir Polizeiwesen, 1988).
Among the critical objects to be protected, there are the river "Reuss", the protected
natural area between the "Reuss delta" and "Polenschachen", the lake "Vierwaldstiitter-
see", and the ground water basin from "Amsteg" to the lake. On the road network of the
Canton Uri, we reckon with one accident every ten years leading to a light pollution of
water surfaces. A more severe contamination of say 1 km of surface water, or ground
water would happen once every 100 years. Note that a weighted surface area equivalent
was used here for comparing the consequences of water contamination by toxic cargo
releases.
The approach followed for this study was as follows. A list of critical substances
was prepared according to the following criteria (TABLE 10.5):
Phenol good I D
Sulfuric Acid very good 1 D
Caustic Soda very good 1 D
Gasoline not soluble 2 D
Acryl nitril not soluble 3 D
Benzene not soluble 3 D
Trichlorethylene bad 3 D
Ethylmercaptan bad 3 D
Potassium cyanide very good 3 I
Ammonium very good 1 I
nitrate
Ethyl-Parathion not soluble 3 I
'" According to HOMMEL "Handhuch tier gefiihrlichen GUIer", Springer Verlag.
Regarding this study it was recognized that 90% of all dangerous goods transported
had to be considered as toxic to water. The distribution of the products into danger
classes was done on the basis of three documents:
• The German catalog of substances potentially dangerous to water ("Katalog der
wassergefiihrdender Stoffe des Bundesministerium des lnnem der BRD ") considers four
toxicity classes (for water):
Class 1: in general not dangerous to water
Class 2: slightly dangerous to water
Class 3: dangerous to water
Class 4: very dangerous to water
The classification is primarily done on the basis of the toxicity to mammalian
animals, toxicity to bacteria, and the toxicity to fishes. This list contains the following
main representatives:
• The BUWAL list of liquids dangerous to water.
• The Swiss classification list for commercial products and substances prepared by the
Fire Prevention Office for Industry and Trade ("Brandverhiitungsdienst Jilr lndustrie
und Gewerbe").
Applying such classification lists to the dangerous goods transported in the Gotthard
region indicates that 85% of all dangerous goods are directly toxic to water (i.e., all
toxic liquids). This corresponds to about 6% of the total transported goods, respectively
to 3.5% of the total number of trucks.
The goods which are indirectly toxic to water include gaseous and solid products
which cannot reach the water directly. About 5% of the dangerous goods and 5% of the
314 CHAPI'ER 10
goods that are not dangerous can however contaminate the water. Altogether this
corresponds to 5% of the total transported goods. respectively 3% of the total number of
trucks.
Transport of dangerous goods through the Gotthard tunnel is subjected to special
regulations and restrictions (TABLE 10.6) regarding the type of dangerous goods which
are allowed to be transported. as well as the quantities. toxicity class. and combined
truck loads (parcel. bulk).
The philosophy followed in this study for estimating accident frequencies that may
lead to contamination of a water bodies was based on two premises:
a) The accident rates of trucks transporting dangerous goods were deduced from the
general trucks accident statistics of the Canton Uri. the Swiss accident statistics. and
some other international statistics. Based on the percentage of trucks transporting
dangerous goods. the accident rates of trucks transporting dangerous cargo was then
estimated.
b) On the other side. the expected number of accidents was deduced directly from the
accident statistics concerning dangerous goods.
Both premises are not totally independent of each other but allow to check up the
results.
TABLE 10.6. Quantity restrictions for special products transiting through the Gotthard
tunnel
Phenol
1 Quantity
(up to)
-
Quantity, with
special permit
-
forbidden
(above)
any quantity
Sulfuric Acid Skg Skg Skg
Caustic Soda 200 kg 200 kg 200 kg
Gasoline SOOkg 500-S.000 kg S.OOOkg
Acryl nitril 10 kg 10 kg 10 kg
Benzene SOOkg 500-S,OOO kg S,OOOkg
Trichlorethylene Skg Skg Skg
EthyJmercaptan - - any quantity
Potassium cyanide - - any quantity
Ammonium 200 kg 200-2,000 kg 2,000 kg
nitrate
Ethyl-Parathion Skg - Skg
What concerns the potential hazards to water bodies, two types of accident scenarios
were considered:
• Accidents leading to a leak involving toxic cargo release (with or without fIre
event) and presenting a direct hazard to water bodies
CASE STUDIES IN REVIEW AND MISCELLANEOUS INFORMATION 315
• Trucks fire incidents, which can cause an indirect hazard to water bodies, if
extinguishing foam/water is used to combat the truck's fires.
According to the Swiss statistics2 concerning the general traffic for the year 1986 to
1989, there were approximately 5,000 accidents per year, of which 40 to 50 accidents
involving trucks. Around half of them, that is 20 to 25 accidents took place on national
roads. Taking into consideration the number of traveled kilometers per truck this
corresponds to an accident rate of 7 x 10 per truck x kIn. In order to calculate the
accident rates for the different road segments it was also necessary to consider the
accidents statistics for personal cars, and on this basis to deduce the segment specific
accident rates for the trucks. These amount to:
20 x 10·' accidents per truck x kID on the "Axenstrasse",
4 x 10·' accidents per truck x kID in the "Seelisberg" tunnel, and
7 x 10-' accidents per truck x kID for the other road segments.
TABlE 10.9. Frequency of occurrence of the accident scenarios for the Canton Uri
The frequency of accidents involving a fire was detennined on the basis of reported
trucks fires (TABLE 10.7). There are no figures available for Switzerland. The figures
were derived by using the data reported by the fire brigade during interventions on the
national roads of the Canton Uri. For the last 10 years period there were only 2-3 trucks
fires reported. About 3% of these trucks were carrying cargoes that were indirectly
dangerous to water bodies. A frequency of 0.08 accidents per year involving trucks fires
CASE STUDIES IN REVIEW AND MISCELLANEOUS INFORMATION 317
can thus be derived (TABLE 10.8). Usually half of these fires are localized to some
parts of the truck and can be extinguished by the truck driver itself. For this study it was
assumed that in 50% of the cases the cargo took fire. According to this premise we find
an frequency of 0.04 accidents per year.
TABLE 10.8 shows the number of transport accidents involving leakage of
dangerous cargo. The compilation was established on the basis of answers to an inquiry
sent to some 18 Swiss Cantons.
Task 1: Collect the following road accidents stotimcs (global traffic accidents, respectively
truck accidents only): Swiss statistics, foreign statistics, and regional stotistics, showing the
number of tonker-trucks accidents per year.
~: Ust all dJJngerous goods transported in the region (type, toxicity class, quantity)
and classify, if possible, the list according to the transit paths.
Task 3: Extract or estimate the Number of Trucks-Accidents (NTA) involving dJJngerous
goods transport, and select all reported accidents involving tonker-truck fire or cargo
leakage, and express the Accident Frequency (AF) on the basis of Traveled Kilometers per
truck (Tr./em), (i.e., AF = NTA ITotol Tr.Km).
Task 4: Using the above accident frequency figures, and /awwing the number per year of
trucks traveling in the region, and the length of the road segments, it is possible to estimate
the number of trucks accidents on the national roads of the region under consideration.
ThllJ.: Select key products according to the BUWAL's list of classified dJJngerous goods.
Classify again and group the dJJngerous goods according to their physical and
toxicological properties in order to obtoin a reduced list corresponding to 10 to 20 key
products toxic to water bodies.
Task 6: Select and list in tobular form all the road segments belonging to a given passage.
Task 7: Indicate also the type of object (i.e., ecosystem, river, lake, ground water, etc.).
Task 8: Estimate the surface area of the object which may reasonably be damaged or
contaminated.
Task 9: Define a weight factor for each object representing the importQnce of the object
with regards to its use (recreation area,jishing, drinking water, game,forest, etc.).
Task 10: Define different scenarios and estimate the accident frequency for each one.
I.-,§!.1l: Estimate for each scenario the potential degree of c:.estruction of the object as a
function of its distance from the road.
Task 12: For each segment multiply the surface area (/em) of the endJJngered object
(belonging to this segment) by the weighting factor (-) and by the degree of destruction (lib)
which has been estimated for the object. The resulting figure (expressed as modified sUrface
area, in /em) represents the extent of the damages for the corresponding segment.
Task 13: Finally add all the modified surface areas belonging to the same type of object to
obtoin the total sUrface damagedlcontominated corresponding to a given class of objects.
Task 14: Establish three or more different scales all of the same length, and place them
above each other on a diagram:
- - one representing the damagedlcontominated surface area (/em) for an ecosystem,
ranging from 0.1 to 10,000 km (on a logarithmic scale).
- - another one for ground water, ranging from 0 to say 4,000 /em (on a logarithmic
scale), etc., and the last onea special linear descriptive scale is used to compare, and
describe the extent of the damages/contamination. This scale is used to obtain the
resulting event severity value, ranging from 0 to 1 (on a linear scale), by projecting the
values (/em) of the previous scales onto it (that is, it relates them to this descriptive
scale). A value of 0.2 corresponds,for example, to an accident; a value of 004 to a big
accident; and a value of 0.5 to 1.0 to a catastrophe.
CASE STUDIES IN REVIEW AND MISCELLANEOUS INFORMATION 319
TABLE 10.11. Matrix representing endangered objects and the degree of destruction/contamination to
be expected for each one
3km
Object 2 Ecosystem 3
6km
In order to calculate the extent of the potential damages, one has to solve the tasks
described in TABLE 10.10. The methodology for assessing the transport risks of road
trucks and road tankers was proposed by EB&P for the Canton Uri (Figure 10.8).
I I I I I I Ewnt aevertIy
I I I I I I Ecoaystem
0.1 1 10 100 1000 10000 km2
I I I I I I Ground water
Figure 10.8. Example of scales used for connecting the different damaged I contaminated
surface areas with the event severity scale.
Federal Department for Roads Construction. The data base shall contain all the
necessary information related to the national roads and their associated maintenance. It
is an integrated management information system using knowledge expert and
ORACLE. The following information shall become available:
• Description of the road network and its particularities
• Description of traffic, environmental factors, and traffic management systems
• Description of the road conditions, state of repair works, etc.
• Air quality measurements data results
• Noise measurements data results
• Accident statistics I Description of
• Land Information System, LIS (INTERGRAPH)
• Heavy transport module
• Expert system with management information and decision support for planing
road repairs and maintenance activities (including GIS and ORACLE).
The system will include decentralized information systems on the basis of
ORACLE, which shall be interconnected over a network.
Several engineering and/or software engineering companies have specialized in the field
of dangerous goods transportation. In the course of the PPR&S(i) we have contacted
and interviewed a few of them, as well as some Cantonal Authorities in charge of
assessing the risks of transportation, in order to gain more insight in the practical
aspects of dangerous goods transportation and risk assessment. It was, however, not
possible to visit all of them since our resources in man power and time'were very much
limited. The companies and authorities mentioned below are just given as example to
show what is being done, in Switzerland, concerning the risk assessment of dangerous
goods transportation.
ASIT Company
Company: AS IT, Aktiengesellschaft flir Sicherheits- und Informationssysteme im
Transportwesen.
Address: ASIT, Thunstrasse 97, CH-3000 Bern, Tel: 031-352 94 77.
Director. Mr. R. Hanni.
(il PPR&S: Poly Project Risk & Safety of technical systems, Swiss Federal Institute of Technology
CASE STUDIES IN REVIEW AND MISCELLANEOUS INFORMATION 321
ASIT is dealing with safety and infonnation systems in the field of transportation.
They are, at present, involved in the preparation of short risk assessment reports (as
required by BUW AL), environmental impact assessment, and proposals for remedial
actions concerning dangerous wastes. It is worth mentioning, that ASIT participated to
the elaboration of BUWAL's Handbook III.
They have developed together with "Springer-Verlag", Heidelberg, the first
European "Safety Infonnation System" (on CD-ROM), dealing with dangerous goods.
This data base gives information on transport classes, physical and toxicological
properties, remedial actions in case of accident, first help measures, etc. The inform-
ation sources were taken from "Hommel Handbook", "Chemdata", "Action Plans of the
Swiss Fire Brigade", "Merck Catalog", "Swiss Poison List" (Schweizer Giftlisten,
BAG), "Goods Catalog of the Swiss Customs Offices", "Safety Data from SUVA"
(Swiss Accident Insurance Institution), "Handbook of the Chemical Industry Compa-
nies" (VCI, Germany).
Another product developed by AS IT, is a "Dangerous Goods Module for Radio-
active Substances", which was developed jointly with the "School for Radiological
Protection" in Villigen, Canton of Aarau. The company ASIT offers also an expert
system to be used in conjunction with their product "Classification of Dangerous Goods
Transports", which integrates the following regulations:
- ADR: European Agreement Concerning the International Transport of Danger-
ous Goods by Road, complemented by SDR and GGVS;
- RID: International Convention Concerning the Transport of Dangerous Goods
by Rail, and RSD and GGVE;
- lATA (DGR): Dangerous Goods Code of lATA (International Air Transportation
Association);
- IMDG: International Maritime Organization;
- P'IT: Al traffic regulations of the Swiss Post-Telephone-Telegraph Authorities.
ASIT has developed different software programs for the Communal and Cantonal
Environmental Protection Services ("Wehrdienste"). These include action plans for
combating oil or chemicals spills, and fires, material management, compatibility lists
for chemicals and materials being stored or transported, alarm plans, cadastral surveys,
etc.
They are suppliers of Swiss national maps (1:25'000) on CD-ROM with an
intelligent program for panning the digitized maps (stored as pixel/vector maps).
ASIT has prepared several reports for assessing the transportation risks:
• Kurzbericht Storfallverordnung (StFV), Teilstrecke N2: Hergiswil-Seelisberg-
tunnel (Kt. Nidwalden), ASIT Nov. 1993.
• Kurzbericht Storfallverordnung (StFV), Teilstrecke Bern-Neuenburg Bahn (BN),
Bemer-Alpenbahn Gesellschaft Bem-Lotschberg-Simplon (BLS), ASIT 15 Juni,
1993.
322 CHAPTER 10
FlNAJOUR Company
Company: FINAJOUR Company in Therwil. BL
Address: Therwil. BL. Tel: 061-721.72.92
Director. Mr. Desboeuss
This software company has developed a computerized railway timetable and is also
offering a computerized global positioning system (GPS) for personal cars, Including
optimal selection of roads. The system is called FINASAT. and makes use of a
Rockwell GPS in connection with an Olivetti Codemo computer (sold for about Fr.
2.500.-). FINAJOUR is presently developing information system for trucking com-
panies dealing with dangerous goods transports.
Other global positioning systems are. for instance. the systems manufactured by
GARMIN (GPS 45. etc.) and the GPS system Pixie from SONY.
Program ALPHA1
Supplier : Rudolf van Loh. Entwicklung und Vertrieb von Software. Cologne,
Germany .
Computer : IBMlDOS-compatible.
Description: The software package ALPHAI ("Ausbildungs und Lem-Programm als
Hilfe fUr Ausbreitungsprognosen") is written in Turbo Pascal and runs on an IBM/AT.
It was developed for use as learning and teaching tool for understanding gas dispersion
and explosion behavior with respect to emergency and rescue actions (i.e., exercising
rescue in case of catastrophes).
Since the mathematical models are very simple. this computer tool cannot be used
for real events but only for training purpose. By using this simple Gauss dispersion
model and other correlation it is possible to consider and make "worst case estimates"
of release and dispersion events on flat terrain (The model does not consider terrain
effects). The approach presented here follows the German directives of "TA-Luft",4 and
VOl directives. s The calculation of BLEVE and associated heat effects is based on
reference6 •
The program allows the calculation of the following events:
• Hot and cold continuous gas emissions
• Cold release (as a result of container failure)
• BLEVE (of known mass)
• BLEVE (of known volume)
Note: BLEVE stands for Boiling Liquid Expanding Vapor Explosion. It is the case
when a container. containing liquefied gas under pressure, is bursting due to violent
boiling as a result of heat effect. In such a case the content of the vessel is released and
vaporizes at once. In case of ignition it bums suddenly. In such a case it is difficult to
predict the time of explosion.
CASE STUDIES IN REVIEW AND MISCELLANEOUS INFORMATION 323
VCE stands for Vapor Cloud Explosion. It concern the escape of a combustible gas
which may ignite within its flammability limits (i.e., 2.2 - 10 vol.-% for propane). If the
gas mixture is too rich, it will only bum and dissipate much heat radiation.
In the case of a BLEVE, heat effects affecting persons or buildings can be considered,
as well as pressure effects resulting from an explosion.
Program PCMLUS-92
Description: The software package PCMLUS-92 was developed by van Loh and
Romberg. 7 It simulates emissions from automotive vehicles along roads with scarce
constructions, with or without vegetation alongside the road. 8 The program can
calculate the following emissions: CO, HC (hydrocarbons), NO, N0 2, Pb, S02' PM
(particulate matter/carbon black), and Benzene. The emissions concentrations can be
calculated on a yearly average or a 98-percentile. The programs assumes different road
conditions like noise protection walls or vegetation and their distance relative to the
road side. Traffic volume, speed and composition are also required.
BL Sichemeitsinspeictorat
4410 Liestal, Rheinstrasse 29
SH Kantonales Laboratorium
8204 Schaffhausen, Postfach 37
The addresses and telephone ,numbers of the official incident reporting offices of the
Cantons (According to article 10, paragraph 2 of the Law for Environmental Protection
("Umweltschutzgesetz"), which have to announce and transmit impressively any
incident to the Federal National Alarm Office (NAZ, "Nationale Alarrnzentrale des
Bundes") are listed in TABLE 10.13 (Status 1992).
TABLE 10.13. Addresses of the centralized Cantonallncident Reporting Offices (for catastrophe pr0-
tection)
BE Einsatzzentraie Kapo
Postfach 300IBem
BL Kantonspolizei
4410 Liestal, Polizeikommando, Mtihlegasse 12114
BS Kantonspolizei Basle-Stadt
4001 Basle, Spiegelgasse 6
GE Commissariat de Police
1211 Geneve 8,17-19, Bd. Carl-Yogt, Case Postale 236
GL Polizeikommando
8750 Glarus
GR Polizeikommando GraubUnden
7000 Chur, Ringstrasse 2
IV Gendarmerie Cantonaie
2800 Delemont, Route de Bille 23
NE Gendarmerie
2000 Neuchiitel, Poudrieres 14
CASE STUDIES IN REVIEW AND MISCELLANEOUS INFORMATION 327
NW Kantonale Polizei
6370 Stans, Kreuzstrasse
OW Kantonale Polizei
6060 Sarnen, Foribachstrasse
SH Kantonspolizei
8200 Schaffhausen, Klosterstrasse 9
SO Einsatzzentrale Kantonspolizei
4502 Solothurn, Ambassadorenhof
SZ Polizeikommando
6430 Schwyz, Bahnhofstrasse 7
UR Kantonales Polizeikommando
6460 Altdorf, Tellsgasse 5
Summary
This chapter includes a short summary of selected case studies and miscellaneous
information of interest to the risk analyst, when dealing with risk assessment of various
transportation technologies and systems. The topics included here are considered as
examples to complement the information presented in the other chapters of this book.
The selection was large in order to include the experience from various countries (e.g.,
328 CHAYIER 10
AC-Laboratory • 241
Actual levels • 266
A Actual loads • 266
Accident Addresses
assessing true accident rates· 58 Swiss Federal and Cantonal Authorities •
failure rate data· 57 324
frequency· 227 Air-mixtures transport • 211
frequency during transportation· 65 Alarm organization
prevention· 222 ARMA, Swiss Meteorological Institute •
propagation· 43 86
reporting· 58; 221 Ammonia· 10
reporting, of· 31 Ammonia pipelines • 190
scenarios· 13 Gulf Central Pipeline· 191
severe, major· 278 Anhydrous ammonia· 197
study on accident type • 31 Annual fatality rates· 160
Accident analysis ARCINFO·4
of rail and road transport risks • 73 Assessment of energy systems· 236
Accident frequency· 306 Average accident rates· 306
estimation, of· 314
involving cargo fire· 316 B
Accident rate
computation, of· 5; 296 Bayes method • 58
estimation for ship transports· 160 BAZL office· 78
railroad • 6; 71 BLEVE·322
road traffic • 5 BLEVE rue ball • 73
Accident scenarios Bulk transport
for ship· 163 by train • 125
for the Canton Uri • 316 liquefied gases • 126
Accident speed liquid chlorine • 126
offreight trains • 130 motor gasoline· 125
Accident statistics of hazardous materials· 158
by accident circumstances· 224 Bulk transportation
by location and dangerous goods class modes, of· 71
(RID). 225 Bulletin
damage to people· 225 accident! incident and inventory· 71
for France· 226 Bunds
for Great-Britain .. 226 for ammonia storage· 196
for Switzerland· 221 purpose, of· 196
for the Netherlands.• 226 types, of· 196
for USA. ·226 Bumingrate
for Western Germany· 223; 224 for liquids· 196
Accidental release BUWAL·121; 136
probability, of· 30 BUWAL Handbook· 10
Accumulation factors· 265 BUWAL Handbook m • 301
Acetylene transport· 211 BUWAL office· 82; 86; 102
332 SUBJECT INDEX