Fire risk analysis of structures and infrastructures: theory and application in highway tunnels

Fire risk analysis of structures and

infrastructures: theory and

application in highway tunnels

CORSO DI PROGETTAZIONE STRUTTURALE ANTINCENDIO, AA: 2015-’16

1November 10 2015www.francobontempi.org

Konstantinos Gkoumas, PhD, PE

Facoltà di Ingegneria

Sapienza Università di Roma

Corso di Progettazione Strutturale Antincendio

Docente: Prof. Ing. Franco Bontempi

http://www.francobontempi.org/

Outline

• System approach to fire safety design

• Risk/fire risk/risk analysis

• Risk assessment process

• Risk analysis

• Hazard analysis

• Risk acceptance

• Risk reduction

• Risk assessment of road tunnels using PIARC/OECD QRAM

• Case study: risk assessment of a long highway tunnel

• References

www.francobontempi.org

CORSO DI PROGETTAZIONE STRUTTURALE ANTINCENDIO

2Fire risk analysis of structures

and infrastructures: theory and









• System approach to fire safety design

• Risk

- fire risk

- risk types

- risk analysis







System approach to fire safety design



MINOR

SPREAD

FIRE SPREADSTOP FIRE

suppressionY

MAJOR

SPREAD

STRUCTURAL

INTEGRITY

AVOID

CASUALITIES

LOCALISED

DAMAGE

STRUCTURAL

FAILURES

N

mitigation

Y

N

fire safe design

Y

N

FIRE

robust design

Y

N

MAJOR

COLLAPSE

AVOID

DIRECT

DAMAGE

AVOID

COLLAPSE

1

2

3

4

0 preventionOBJECTIVE

fire safety design -

structural

fire safety design -

non structural

GLOBAL

SAFETY

LOSS OF

GLOBAL

SAFETY

AVOID

INDIRECT

DAMAGE

NY

The fire safety is framed in different

“safety levels”, corresponding to

different safety objectives.





(Fire) Risk Estimation*

*(following SFPE Handbook of Fire Protection Engineering)

Provide answer to the following questions

1. What could happen?

2. How bad would it be if it did happen?

3. How likely is it to happen







What is risk?

Risk can be defined as the probability that the harm or damage from a particular

hazard is realized.

Risk is measured in terms of consequences and likelihood (a qualitative description

of probability or frequency). In mathematical terms risk can be defined as:

risk = f (frequency or probability, consequence) (1)

In the case of an activity with only one event with potential consequences, a risk (R)

is the probability (P) that this event will occur multiplied with the consequences (C)

given the event occurs:

R = PC (2)

The risk of a system is the sum of the risks of all harmful events of that system:

(3)



𝑅𝑆 = 𝑅𝑖

𝑛

𝑖=1





Risk types• Life safety risks are normally presented in two ways:

- Individual risk and

- Societal risk

• Individual risk:

The purpose of the individual risk is to ensure that individuals in the society

are not exposed to unacceptably high risks. It can be defined as the risk to any

occupant on the scene for the event/hazard scenario i.e. it is the risk to an

individual and not to a group of people.

• Societal risk:

Societal risk is not looking at one individual but is concerned with the risk of

multiple fatalities. People are treated as a group, there are no considerations

taken to the individuals within the group i.e. the definition of the risk is from a

societal point of view.



Source: Jönsson, 2007





What is risk analysis?

• A big family of different approaches, methods

and complex models combining various

methododical components for specific tasks

• Systematic analysis of sequences and interaction

effects in potential accidents, thereby identifying

weak points in the system and recognizing

possible improvement measures

• Risk analysis makes the quantification of risks

feasible







The risk assessment process







The risk assessment process



Start

Definition of the system

Hazard identification

Probability analysis Consequence analysis

Additional safety

measures

Risk estimation

Risk evaluation Risk criteria

Acceptable

risk?

Stop

Risk analysis

Risk evaluation

YES

NO

Risk reduction





Definition of the system (context establishment)



Define the operational environment and the context of the risk assessment

process

– Definition of the scope or the risk assessment process

• This includes determining the timeframe (e.g. from planning to dismantling), the

required resources and the depth of analysis required.

– Definition of the strategic and organizational context

• The nature of the organization in charge of the risk management and the

environment in which it operates is established

– Identification of the stakeholders (portatori di interesse) and objectives

• The relationships that are interdependent with the organization are defined, the

impacts that might occur are accounted for, as well as and what each is wanting

out of the relationship

– Determination of the evaluation criteria

• Decide what level of risk the organization is prepared to accept









a. What can happen

b. How can it happen

Means for hazard identification:

• Decomposition of the system into a number of

components and/or subsystems

• Identification of possible states of failure for the

considered system and sub-systems

• Identification of how the hazards might be realized

for the considered system and subsystemsSource: Faber, 2008







Hazard identification – system decomposition

A. Structure

1. Main components

(d) Foundations

(c) Towers

(b) Anchor systems

(a) Main cables

(h) Cable saddle

(e) Railway girder

(f) Highway girders

(g) Expansion joints

(e) Non str.elements

(a) Steel

(b) Concrete

(c) Prestressed c.

(d) Alluminium/iron

3. Materials

(f) Coating

4. Systems

(a) Electrical

(c) Hydraulics

(b) Mechanical

(e) Bitumen

(e) Plastic

2. Secondary comp.

(d) H.R. attachments

(c) TMD

(b) Buffers

(a) Hanger ropes

B. Users

1. Highway traffic

(b) Commercial

(a) Private

2. Railway traffic

(b) Commercial

(a) Private

(a) Heavy

(b) Hazard mat.

(c) Military

3. Exceptional traffic

C. Facilities

1. Over the bridge

(b) Railway

(a) Highway

2. By the bridge

(a) Highway

(b) Railway

(c) Toll booths

(d) Control center

(e) Parking

(a) Maritime traffic

3. Under the bridge

D. Dependencies

1. Power

3. Financial

2. Communications

4. Supplies

5. Emerg. Responce

(a) First aid

(b) Police

(c) Fire brigade

(d) Hospitals

6. Ext. Contractors

E. Linkage

1. Economy

3. Military

2. Social

F. Operation

1. Authorities

(b) Management

2. Aspects

(a) Bridge authorities

(b) Goverment

(c) Region

5. Personnel

(c) Maintenance

(a) Financial

(b) Other

(a) Technical

G. Technology

(a) GPS

(b) Accelerometers

(c) Strain gauges

(e) Thermometers

(g) CCTV

(f) WIM

(d) Seismographs

(h) Field equipment

1. Monitoring

2. Control

(a) Cable control

(d) Railway traffic

(c) Highway traffic

(b) TMD

3. Data transmission

(b) Wireless

(a) Cable

4. Computer center

(b) Software

(a) Hardware

(d) Internet/LAN

(c) Data bases

4. Regulations

3. Policies

4. Location

(c) External

Hie

rarc

hic

al

Ho

log

rap

hic

Mod

els

(HH

M)

(Def

ined

in H

aim

es,

19

81

)





Risk analysis: hazard identification



• Qualitative methods

Studies based on the generic experience of personnel and do not

involve mathematical estimations.

• Quantitative methods

Mathematical estimations that rely upon historical evidence or

estimates of failures to predict the occurrence of an event.

• Semi-quantitative methods

Combination of the above (mostly, qualitative methods with

applied numerical values).

Source: Nolan, D. P. Handbook of Fire and Explosion Protection

Engineering Principles for Oil, Gas, Chemical, and Related Facilities, 1986







Source: Aven, T. Risk Analysis: Assessing Uncertainties beyond Expected Values and Probabilities.

John Wiley & Sons, 2008

Risk analysis: hazard identification





Hazard identification. Qualitative Methods



Checklist or Worksheet

A standardized listing which identifies common protection features required for typical

facilities is compared against the facility design and operation. Risks are expressed by

the omission of safety systems or system features.

Preliminary Hazard Analysis (PHA)

Each hazard is identified with potential causes and effects. Recommendations or known

protective measures are listed.

What-If analysis

A safety study which by which “What-If’ investigative questions (brainstorming

approach) are asked by an experienced team of a hydrocarbon system or components

under examination. Risks are normally expressed in a qualitative numerical series (e.g., 1

to 5).

HAZOP - HAZard and OPerability analysis (analisi di pericolo e operabilità)

A formal systematic critical safety study where deviations of design intent of each

component are formulated and analyzed from a standardized list. Risks are typically

expressed in a qualitative numerical series (e.g., 1 to 5) relative to one another.

Source: Nolan, D.P. 1986. Handbook of Fire and Explosion Protection Engineering Principles for …. Noyes, New Jersey





Hazard identification. Qualitative Methods



Event Trees (ET) –albero degli eventi

A mathematical logic model that mathematically and graphically

portrays the combination of events and circumstances in an

accident sequence, expressed in an annual estimation.

Fault Trees (FT) – alberi dei guasti

A mathematical logic model that mathematically and graphically

portrays the combination of failures that can lead to a specific main

failure or accident of interest, expressed in an annual estimation.

Failure Modes and Effects Analysis (FMEA)

A systematic, tabular method of evaluating the causes and effects

of known types of component failures, expressed in an annual

estimation.

Source: Nolan, D.P. 1986. Handbook of Fire and Explosion Protection Engineering Principles for …. Noyes, New Jersey





• Risk analysis

• Qualitative risk analysis

• Quantitative risk analysis









Risk analysis

• Risk analysis

– Probability- as the likelihood of the risk occurrence

– Impact - consequences if the risk occurs

• risk proximity, meant as the point in time during which

a risk will impact

• Risk analysis - methods

– Qualitative Risk Analysis, in which numbers and

probabilities are used not extensively or at all

– Quantified Risk Analysis (QRA)

– Probabilistic Risk Analysis (PRA), in which the system risk

is represented as a probability distribution







Risk analysis and system complexity

High-Probability/

Low-Consequences

(HPLC)

Low-Probability/

High-Consequences

(LPHC)

High-Probability/

Low-Consequences

(HPLC)

Low-Probability/

High-Consequences

(LPHC)

High-Probability/

Low-Consequences

(HPLC)

Low-Probability/

High-Consequences

(LPHC)

High-Probability/

Low-Consequences

(HPLC)

Stochastic

Complexity

Deterministic

Analysis

Methods

Qualitative

Risk

Analysis

Quantitative/Probabilistic

Risk

Analysis

Pragmatic

Risk

Scenarios

Stochastic

Complexity

Deterministic

Analysis

Methods

Qualitative

Risk

Analysis

Quantitative/Probabilistic

Risk

Analysis

Pragmatic

Risk

Scenarios







Qualitative Risk analysis

• Qualitative Risk Analysis is the simplest method of risk analysis, and

generally is used during the preliminary analysis phases.

• It consists in using subjective assessments of risks, and consequently, in

ranking them in a subjective manner.

• Sources for information to be used in the analysis can be drown from

previous experiences, history of events and consultation of experts.

• The ranking of risks is qualitative, e.g. risk (1) > risk (2) > risk (3),

while a description can be added. Eventually, a likelihood-consequence

matrix can be constructed.

• The biggest drawback of QRA is that there is neither a clear indication

of the risk’s magnitude nor an absolute scale of how serious the risk

might be, so, for a comprehensive risk analysis of more complex

systems, quantitative methods should be preferred.







Qualitative risk analysis methods: risk matrix• A risk matrix typically provides a discrete partitioning of relative consequences

along one dimension and relative likelihood along the other.

• The entry in each matrix cell may include a description of hazards known or

believed to have that combination of consequence severity and likelihood.

Source: NFPA, SFPE Handbook of

Fire Protection Engineering,

3rd edition, 2002

Source: Furness, A., Muckett, M.

Introduction to Fire Safety

Management. Elsevier, 2007.







Qualitative risk analysis methods: SWOT analysis

24

Strengths (forza): characteristics of

the business or project that give it

an advantage over others.

Weaknesses (debolezza):

characteristics that place the

business or project at a

disadvantage relative to others

Opportunities (opportunità):

elements that the project could

exploit to its advantage

Threats (minacce): elements in the

environment that could cause

trouble for the business or project

Fire risk analysis of structures






Quantitative Risk analysis

• Quantified (or quantitative) Risk Analysis (QRA) combines

the consequences and frequencies of accident scenarios to

estimate the level of risk.

• In respect to the Qualitative method, QRA implicates the

acquaintance of probabilities that describe the likelihood of

the outcomes and their consequences.

• QRA started with the chemical industries from the 70s and

the offshore industry from the 80s.

• QRA is traditionally expressed through the decomposition

of the system. This frequently is done by the use of event

trees and fault trees.







FTA and ETA

• ETA (event tree analysis – albero degli eventi)

provides a structure for postulating an initiating

event and analyzing the potential outcomes

• FTA (fault tree analysis – albero dei guasti)

begins with a failure and provides a structure to

look for potential causes







Event tree analysis

• Event trees pictorially represent the logical order in which

events in a system can occur. Event trees begin with an

initiating event, and then the consequences of the event are

followed through a series of possible paths.

• Each path is assigned a probability of occurrence. Therefore,

the probability of the various possible outcomes can be

calculated.

• Event tree analysis is based on binary logic, in which an

event has either happened or not, or a component has failed

or has not.

• It is valuable to analyze the consequences arising from a

failure or undesired event.







Event tree analysis: illustration (1)

28

Event trees are helpful in

considering all the possible

outcomes (on the right-hand side)

from an initiating event (on the

left-hand side), which is usually

ignition for fire risks.

The frequency of the initiating

event can be estimated from fire

report data, and the conditional

probabilities of the sub-events can

be quantified from fire report data

or fault trees.







Event tree analysis: illustration (2)

Source: Fire Risk in Metro Tunnels and Stations Hyder Consulting







Fault tree analysis

30

Fault trees are helpful in

quantifying the

probability of a top

event of concern (such

as the failure of a fire

protection system) from

all the potential root

causes (at the bottom),

again quantified from fire

report data.







Fault tree analysisgeneral conclusion (event)

• Fault trees look like a complement

to event trees.

• The idea is to begin with a general

conclusion (event) and, using a

top-down approach, to generate a

logic model that provides for both

qualitative and quantitative

evaluation of the system

reliability.

Source: google pictures search “Fault tree”







Fault tree analysis - symbolsBasic event - failure or error in a system component or element

(example: switch stuck in open position)

Initiating event - an external event (example: bird strike to aircraft)

Undeveloped event - an event about which insufficient information is

available, or which is of no consequence

Conditioning event - conditions that restrict or affect logic gates

(example: mode of operation in effect)

Intermediate event: can be used immediately above a primary event to

provide more room to type the event description.

Source: Fault Tree Handbook. Nuclear Regulatory Commission. NUREG–0492







Fault tree analysis – gate symbols

OR gate - the output occurs if any input occurs

AND gate - the output occurs only if all inputs occur (inputs are

independent)

Exclusive OR gate - the output occurs if exactly one input occurs

Priority AND gate - the output occurs if the inputs occur in a specific

sequence specified by a conditioning event

Inhibit gate - the output occurs if the input occurs under an enabling

condition specified by a conditioning eventSource: Fault Tree Handbook. Nuclear Regulatory Commission. NUREG–0492







Advantages and disadvantages of FTA

• Disadvantages

1. There is a possibility of oversight and omission of significant failure

modes.

2. It is difficult to apply Boolean logic to describe failures of system

components that can be partially successful in operation and thereby

affect the operation of the system, e.g. leakage through a valve.

3. For the quantitative analysis there is usually a lack of pertinent failure

data. Even when there are data they may have been obtained from a

different environment.

• Advantages

1. It provides a systematic procedure for identifying faults that can exist

within a system.

2. It forces the analyst to understand the system thoroughly.Source: Hasofer et al. 2007, Risk Analysis in Building Fire Safety Engineering







Cause – consequence diagrams

• The combination of fault trees and event trees leads to the creation of

cause-consequence diagrams.

Time

Revealed from the

Monitoring system

S3

S2

S1

Consequences

Infraction of traffic law

Improper speed

Road condition

Vehicle flow

blocked

YES

YES

NO

NO

Other

Iniziative event

Road

Accident







SCENARIO PROBABILITY

A1 PA*P1

A2 PA*(1-P1) *P2 *P3

A3 PA*(1-P1) *P2*(1-P3 )

A4 PA*(1-P1) *(1-P2)*P3

A5 PA*(1-P1) *(1-P2)*(1-P3)

B1PB*P1

B2PB*(1-P1) *P2 *P3

B3PB*(1-P1) *P2*(1-P3 )

B4PB*(1-P1) *(1-P2)*P3

B5 PB*(1-P1) *(1-P2)*(1-P3)

C1PC*P1

C2PC*(1-P1) *P2 *P3

C3PC*(1-P1) *P2*(1-P3 )

C4PC*(1-P1) *(1-P2)*P3

C5 PC*(1-P1) *(1-P2)*(1-P3)

Triggering

event

Fire

ignition

1. Fire

extinguished

by personnel

2. Intrusion of

fire fighters

Arson

Explosion

Short

circuit

Cigarette

fire

YES (P1)

NO (1-P1)YES (P2)

NO (1-P2)

Scenario

Other

A1

A2

A3

A4

A5

3. Fire

suppression

YES (P3)NO (1-P3)

YES (P3)NO (1-P3)

Fire

location

AREA A

(PA)

YES (P1)

NO (1-P1) YES (P2)

NO (1-P2)

B1

B2

B3

B4

B5

YES (P3)NO (1-P3)

YES (P3)NO (1-P3)

AREA B

(PB)

YES (P1)

NO (1-P1) YES (P2)

NO (1-P2)

C1

C2

C3

C4

C5

YES (P3)NO (1-P3)

YES (P3)NO (1-P3)

AREA C

(PC)

Quantified Risk Analysis: cause – effect diagrams







F (frequency) – N (number of fatalities) curve

• An F–N curve is an alternative way of describing the

risk associated with loss of lives.

• An F–N curve shows the frequency (i.e. the expected

number) of accident events with at least N fatalities,

where the axes normally are Logarithmic.

• The F–N curve describes risk related to large-scale

accidents, and is thus especially suited for

characterizing societal risk.

Source: Aven, T. Risk Analysis: Assessing Uncertainties beyond Expected Values and Probabilities. John Wiley & Sons, 2008








Source: Aven, T. Risk Analysis: Assessing Uncertainties beyond Expected Values and Probabilities. John Wiley & Sons, 2008








Source: NFPA, SFPE Handbook of Fire Protection Engineering, 3rd edition, 2002





Index

• Risk acceptance

- ALARP

- Human life (!)









Risk acceptance

Source: Persson, M. Quantitative Risk Analysis Procedure for the Fire Evacuation of a Road Tunnel -An Illustrative Example. Lund, 2002







Risk acceptance – ALARP (1)

RISK MAGNITUDE

INTOLERABLE

REGION

As

Low

As

Reasonably

Practicable

BROADLY ACCEPTABLE

REGION

Risk cannot be justified

in any circumstances

Tolerable only if risk

reduction is impracticable

or if its cost is greatly

disproportionate to the

improvement gained

Tolerable if cost of

reduction would exceed

the improvements gained

Necessary to maintain

assurance that the risk

remains at this level

As

Low

As

Reasonably

Achievable

RISK MAGNITUDE

INTOLERABLE

REGION

As

Low

As

Reasonably

Practicable

BROADLY ACCEPTABLE

REGION

Risk cannot be justified

in any circumstances

Tolerable only if risk

reduction is impracticable

or if its cost is greatly

disproportionate to the

improvement gained

Tolerable if cost of

reduction would exceed

the improvements gained

Necessary to maintain

assurance that the risk

remains at this level

As

Low

As

Reasonably

Achievable







Risk acceptance – ALARP (2)

Source: google pictures search “ALARP”






Monetary values – cost of human life (!)

• What is the maximum amount the society (or the decision-maker) is willing

to pay to reduce the expected number of fatalities by 1?

• Typical numbers for the value of a statistical life used in cost-benefit analysis

are 1–10 million euros. The Ministry of Finance in Norway has arrived at a

value at approximately 2 million euros.


Guideline values for the cost to

avert a statistical life (euros), used

by an oil company

Source: Aven, T. Risk Analysis: Assessing

Uncertainties beyond Expected Values and

Probabilities. John Wiley & Sons, 2008





Risk reduction









Risk reduction

Source: Brussaard et al. 2004. The Dutch Model for the Quantitative Risk Analysis of Road Tunnels.







Risk reduction (2) - monitoring and system response

Time

1

3

2

Accident Accident evolutionPre-accident

situation

Pre-accident

Monitoring

Pre-accident

System Response

Accident

Localization

Evolution of System Response

Accident evolution Monitoring

System

Response







48

Parte 2a: Analisi del rischio di gallerie stradali




Source:

"Risk analysis for severe traffic accidents in road tunnels". “Laurea Magistrale”

(M.Sc.) Thesis at the Sapienza University of Rome, Faculty of Civil and Industrial

Engineering. Candidate: Carmine Di Santo. Final grade: 110/110 “Summa cum

Laude”. Advisor: Prof. Franco Bontempi, co-advisor: Konstantinos Gkoumas, PhD.

Defended in January 2015.


www.francobontempi.org 49

The issue of SAFETY in tunnels

Mont Blanc Tunnel Fire (1999)

39 Fatalities

Italia (Courmayer) – France (Chamonix)

single – bore, bidirectional tunnel

Length = 11.6 km

St. Gotthard Tunnel Fire (2001)

11 Fatalities

Switzerland (Göschenen) – Switzerland (Airolo)


Length = 16,9 km

Frejus Tunnel Fire (2005)

2 Fatalities

Italia (Bardonecchia) – France (Modane)


Length = 12,9 km






The issue of SAFETY in tunnels

Directive

2004/54/ECQuantitative Risk

Analysis

Objectives

Parameters

Requirements

Transport of Dangerous Goods through road tunnels

OECD/PIARC/EU

Quantitative Risk Assessment Model

• OECD (Organisation of Economic Co-

operation and Development)

• PIARC (World Road Association)

• European Commission

• France (INERIS), Canada (WS Atkins),

UK (IRR)






PIARC/OECD QRAM OUTPUTS

𝑅 = 𝐹 ∙ 𝐶F probability of occurrence / frequency

C extent of damage / consequences

• Fatalities

• Injured

• Destruction of buildings and structures

• Environmental Damage

PIARC/OECD

QRAMSocietal Risk

Individual Risk F

[1/y

ear]

N [Fat]

𝐸𝑉𝑠 =

𝑖=0

∞

𝐹 𝑁𝑖 ∙ 𝑁𝑖

The risk to which a group of people

is subjected in case a scenario s

occurs.

Prob. that a person (among local

population and within a certain

distance from the road) dies due to

the scenario s.






PIARC/OECD QRAM OUTPUTS

𝑅 = 𝐹 ∙ 𝐶F probability of occurrence / frequency

C extent of damage / consequences

• Fatalities

• Injured

• Destruction of buildings and structures

• Environmental DamagePIARC/OECD

QRAMSocietal Risk

F [1

/yea

r]

N [Fat]

𝐸𝑉𝑠 =

𝑖=0

∞

𝐹 𝑁𝑖 ∙ 𝑁𝑖

The risk to which a group of

people is subjected in case a

scenario s occurs.

SR = F(N) ∙ N

The F-N diagrams may be

applied to illustrate the risk

profile for a specific hazard

such as a fire in a road tunnel.

Expected amount of victims

in a certain time period






F-N curve construction process flow

1 – Dangerous Goods and

Accident Scenarios selection

2 – Effect j (due to the scenario s ) and its Range from the epicentre

Ej=f(d) Rj

3 – Mortality Rate within the range Rj %LETHj

4 – Mortality Rate corrected considering the possibility of escape

%LETHj=f(tevac)

5 – Scenario s Probability of

occurrence fs

6 – Number of victims due to the scenario s

N= jNj=f(Rj, Dru, Ljam, %LETHj)

7 – F-N curve for the scenario s and its relative Expected Value






Tunnel risk assessment procedure

Data Collection

Data Preparation

Risk Calculation

Using QRAM

Is Risk

acceptable?

NO

Additional risk

reduction

measures

YES

End

Mean Data:

• Traffic

• Accident Frequencies

• Tunnel Geometry

• Tunnel Equipment

• .....

Risk Acceptability:

• Absolute criteria

• Relative criteria

Risk

Analysis

Prevention Measures:• Signs and road markings

• Lighting

• Traffic control

• Route geometry

• Prohibition of access to

certain types of vehicles

Protection Measures:• Monitoring

• Fires/Accident Detection

system

• Ventilation system

• Emergency lighting

• Protection of escape routes

• System of emergency

management

• Emergency Procedures






Societal Risk Acceptability criteria

Absolute

Criteria𝐸𝑉𝑠 ≤ 𝐸𝑉𝑙𝑖𝑚𝑖𝑡 (𝐸𝑉𝑠: Expected Value of Victims)

0,000001

0,00001

0,0001

0,001

0,01

0,1

1

1 10 100

F(N

) [1

/ye

ar]

N

Tollerable Risk Line Acceptable Risk Line

Not Acceptable area

ALARP

Acceptable area

As Low As Reasonably Practicable

ALARP area:

• prevention and/or

mitigation actions must be

taken to reduce the risk, as

far as reasonably

practicable

• Cost – Benefit Analysis






Societal Risk Acceptability criteria

Relative Criteria 𝐸𝑉𝑠 ≤ 𝐸𝑉𝑠,𝑟𝑒𝑓

1,00E-06

1,00E-05

1,00E-04

1,00E-03

1,00E-02

1,00E-01

1,00E+00

1,00 10,00 100,00 1000,00

Tunnel Tunnel Reference

Applying the same calculation method, compare the examined risk with:

• the risk of an alternative route

• that calculated for a reference tunnel, which must have characteristics similar to the one

examined, but with all the safety requirements required by the relevant regulations






Probit analysis

Is a type of regression used to analysing the relationsheep between a stimulus

(dose) and “all or nothing” (such as death) response

The following items must be identified:

• The toxicant

• The target

• The effect or response to be monitored

• The dose range

• The period of the test

Biological organisms respond differently

to the same dose of a toxicant.

Each individual is exposed to the same

dose and the response is recorded.

Curves are frequently represented by a

normal or Gaussian distribution

A Gaussian or normal distribution

representing the biological response to

exposure to a toxicant.

𝑓 𝑥 =1

𝜎 2𝜋𝑒−𝑥−𝜇 2

2𝜎2

probability (or fraction) of individuals

experiencing a specific response

x is the response, σ is the standard

deviation, and μ is the mean.

σ determines the shape and μ characterize the location of the curve with respect to the x axis

the percentage

of individuals

affected for a

specified

response interval

• The toxicological experiment is repeated for a number

of different doses, and normal curves are drawn.

• The standard deviation and mean response are

determined from the data for each dose.

FINNEY 1971






Probit analysis

A complete dose-response curve is produced

by plotting the cumulative mean response at

each dose.

The response is plotted versus the logarithm of the

dose, to provide a much straighter line in the

middle of the response curve

For comparison purposes the dose that results

in 50% lethality of the subjects is frequently

reported. This is called the LD50 dose

(lethal dose for 50% of the subjects).

For computational purposes the response

versus dose curve is not convenient.

For single exposures the probit method is

particularly suited, providing a straight-line

equivalent to the response-dose curve.

P or RATIO =1

2𝜋

−∞

Pr−5

𝑒−12𝑢2𝑑𝑢

provides a relationship between the

probability P and the probit variable Pr.

Many methods exist for representing the

response-dose curve.






Probit analysis

Transformation from Percentages to Probits

The probit relationship transforms the sigmoid shape of the normal response versus

dose curve into a straight line when plotted using a linear probit scale

The probit variable Pr is computed from

𝑃𝑟 = 𝑎 + 𝑏 ln 𝐷

P or RATIO =1

2𝜋

−∞

Pr−5







1) Dangerous goods and accident scenarios

Liquefied

Petroleum Gas

(LPG)

Motor Spirit

Acrolein

(Toxic Liquid)

Chlorine

(Toxic Gas)

Ammonia

(Toxic Gas)

Liquified CO2

Boiling Liquid Expanding

Vapor Explosion (BLEVE)

Toxic Release in the air

Torch Fire

Pool Fire

Vapor Cloud Explosion (VCE)

Vapor Cloud Explosion

BLEVE

No DGs 20MW Fire

100 MW Fire






0

20

40

60

80

100

120

140

0,00 10,00 20,00 30,00 40,00 50,00 60,00 70,00

qr

[kW

/m2]

d [m]

2) scenario physical effects

Thermal Effects Pressure Effects Toxicity Effects

• Fires

• VCEs

• BLEVEs

• VCEs

• BLEVEs

• Fires (smokes)

• Toxis Releases in

air

qr [kW/m2] = f(d)

Radiative Heat Flux

which is experienced

by the receiver per

unit area

Side-on Blast Overpressure

∆Ps [bar] = f(d)

Wave Positive-phase

t+[bar] = f(d)

Concentration

C [ppmv] = f(d)

Effect

Intensity

Distance from

the epicentre






3) Physiological effects𝐹𝑖𝑛𝑛𝑒𝑦,1971

𝑅𝑎𝑡𝑖𝑜 = 𝑓(Pr) =

−∞

𝑃𝑟−5


𝑃𝑟𝑜𝑏𝑖𝑡 𝑇𝑟𝑎𝑛𝑠𝑓𝑜𝑟𝑚𝑎𝑡𝑖𝑜𝑛

𝑃𝑟𝑗 = 𝑎 + 𝑏 ln 𝐸𝑗 ∙ 𝑡𝑒𝑥𝑝,𝑗

Thermal

PressureToxicity






4) probability of occurrence of the scenarios

𝑓𝑠,𝑖 = 𝑃𝑠,𝑖 ∙ 𝑓𝑎𝑐𝑐,𝑖 ∙ 𝑇𝐻𝑖 ∙ 𝐿𝑖 ∙ 24 ∙ 365 ∙ 10−6

Frequency of occurrence of the scenario s on the section i in a year [scen/year]

Conditional probability

that scenario j occurs

once an accident

implying an HGV has

taken place on the

section i

Annual frequency of accidents involving HGVs on the section i

[acc/(MVkm*year)]

Traffic of HGVs passing through the section i in one hour

[veh/h]

𝑓𝑎𝑐𝑐,𝑖 =𝐻𝐺𝑉𝑎𝑐𝑐,𝑖𝑇𝐻𝑖 ∙ 𝐿𝑖

Fault T

ree A

naly

sis

• HGV/h

• % DG-HGV

• DGs types

• Accidents/year

𝐏 = 𝐏𝟏. 𝟏 ∙ 𝐏𝟏. 𝟐 + 𝐏𝟐. 𝟏 ∙ 𝐏𝟐. 𝟐






5) Possibility of escape or of finding shelter

𝐷𝑖𝑗 = 𝑡𝑖𝑛

𝑡𝑜𝑢𝑡

𝐷𝑗 𝑡 𝑑𝑡

Dose of physical effect j

that affects a man

crossing the segment i

𝐷𝑗,𝑇𝑂𝑇 =

𝑖

𝐷𝑖𝑗

Total dose

received during

the escape

𝑃𝑟𝑗 = 𝑎 + 𝑏 ∙ ln 𝐸𝑗 ∙ 𝑡𝑒𝑥𝑝,𝑗

𝐷𝑗,𝑇𝑂𝑇

𝑡𝑒𝑣𝑎𝑐

𝑡𝑝𝑟𝑒 𝑡𝑚𝑜𝑣Pre-movement

timeMovement

time

𝑣

𝑑𝑠𝑎𝑓𝑒𝑡𝑦

𝑡𝑝𝑟𝑒 = 𝑡𝑝𝑟𝑒−𝑏𝑝𝑠 ∙ 𝑤𝑒𝑓𝑓

𝑤𝑒𝑓𝑓 =5

𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑟𝑎𝑡𝑖𝑛𝑔 𝑜𝑓 𝑝𝑎𝑟𝑎𝑚𝑒𝑡𝑒𝑟𝑠

• Alertness (4)

• Mobility (4)

• Social Affiliation (3)

• Commitment (3)

• Familiarity (2)

• Distance from the accident (by calc)

• Perceived severity (4)

Occupant

Response Model

𝑡𝑟𝑒𝑐 + 𝑡𝑟𝑒𝑠






6) Societal risk indicators

Number of Victims

tsce Occurs the accident

scenario

tbarr Delay for stopping

approaching traffic

tjam min (tsce, tbarr).

N = 𝑅 ∙ 𝐷𝑅𝑈𝐽 + 𝑅 − 𝐿𝑗𝑎𝑚 ∙ 𝐷𝑅𝑈𝐹 ∙ %𝐿𝐸𝑇𝐻

Road Users Density in a Traffic Jam [users/m]

Road Users Density in a Fluid Traffic [users/m]






7) Societal risk indicators

F-N curve construction

Each scenario s may appear as different events Ei depending on:

the section of the path being considered (section i)

the accident location on the section

the traffic direction (A, B)

the reference period of the day (QUIET, NORMAL, PEAK)

....

Event Event Frequency Fatalities Cumulative Frequency

Ei fi Ni Fi

[-] [1/year] [Fat] [1/year]

E1 f1 N1 F1 = f1

E2 f2 N2 F2 = f1+f2

E3 f3 N3 F3 = f1+f2+f3

E4 f4 N4 F4 = f1+f2+f3+f4

... ... ... ...

En fn Nn Fn = f1+f2+f3+f4+...+fn

Scenario "s"

F [1

/yea

r]N [Fat]

𝐸𝑉𝑠 = 1

+∞

)𝐹(𝑁 𝑑𝑁






Parte 2b: Analisi del rischio della galleria St. Demetrio






Tunnel St. Demetrio

Motorway Catania – Syracuse

(European route E45)

ANAS s.p.a.

Construction:

2007-2009

Pizzarotti & C. S.p.A. Parma

Courtesy of Dr. Luigi Carrarini

(ANAS S.p.A.)

Courtesy of Ing. Alessandra Lo Cane

(M.I.T.)






Tunnel St. Demetrio

(Central Design Management ANAS S.p.A.)

Height from the Roadway to the Inner Wall 8.06 [m]

Road Platform Width 11.2 [m]

Cross Sectional Area 87.31 [m2]

Natural Tunnel

TWIN BORE TUNNEL,

ONE DIRECTION PER BORE

Polycentric Circular Section

Traditional Excavation

Bore in direction SOUTH (Syracuse)

portal of entry [km] 4+800

portal altitude above sea level [m] 10642

portal of exit [km] 7+695


Length [km] 2895

maximum longitudinal slope [%] 0.32

minimum longitudinal slope [%] 0.32

average longitudinal slope [%] 0.32

Bore in direction NORTH (Catania)

portal of entry [km] 7+698


portal of exit [km] 4+750


Length [km] 2949

maximum longitudinal slope [%] -0.32

minimum longitudinal slope [%] -0.32

average longitudinal slope [%] -0.32

Catania – Syracuse (E45), ANAS s.p.a.

2007-2009, Pizzarotti & C. S.p.A. Parma






Tunnel St. Demetrio: equipment & traffic data

• Pedestrian Bypass every 300m

• Bypass Carriageable every 900m

• Control Centre → Catania

• CCTV cameras placed every 282m

• CO sensors

• Smoke Meters (Opacimeters)

• Linear Thermal Sensors (heat sensing

cable)

• Variable Message Panels every 300m

• SOS stations every 200m

9 Jet Fan

10 Jet Fan 9 Jet Fan

9 Jet Fan

Equipment Emergency Ventilation System

Longitudinal Ventilation

average speed (on the cross section) of

3 m/s in the direction of traffic

time of fire detection (via thermo sensitive

cable) of 3 minutes from the ignition

a time of 5 minutes for the emergency

ventilation establishment






Tunnel St. Demetrio: equipment & traffic data

Traffic

QUIET NORMAL PEAK

AADT = 21190 veh/day

(1 direction)

22÷07

325 veh/h

HGV-ratio = 2%

vcar = 126.4 km/h

vHGV = 90.5 km/h

1050 veh/h

HGV-ratio = 10%

vcar = 126.4 km/h

vHGV = 90.5 km/h

1553 veh/h

HGV-ratio = 11.7%

vcar = 114.5 km/h

vHGV = 82 km/h

SOUTH NORTH

QUIET 1 1

NORMAL 7 3

PEAK 12 5

DG-HGV / h

63% Flammable Liquids (motor spirit, diesel oil, etc.)

31% LPG

6% Others

[acc /(MVkm*year)] [acc /(veh*km*year)]

SOUTH 0.161 0.000000161

NORTH 0.160 0.000000160facc






21 - TUNNEL 3

2 3 - TUNNEL1

SOUTH (Syracuse)

NORTH (Catania)

x

Tunnel St. Demetrio: QRAM input data

Accident Scenarios

Tunnel

QRAM Model

Average number of people in a light vehicle [-] 2

Average number of people in a HGV [-] 1.1

Average number of people in a Bus/Coach [-] 40

Bus/Coaches ratio [-] 0.01

Delay for stopping approaching traffic [s] 9000

Area (Urban/Rural) [-] urban

Average density of population [hab/km2] 0.01

DG transport correction factor [-] 1.00E+00

Traffic & Population Data

W (effective width) [m] 11

H (effective height) [m] 7.9

A (open cross sectional area) [m2] 86.9

Cam (camber) [%] 0

Gs (Segment gradient) [%] 0.32

VnN (volume flow rate along tunnel at nodes) [m3/s] 0

VnE (volume flow rate along tunnel at nodes) [m3/s] 261

Ad (open area of discrete drains) [m2] 0.075

Xd (interval between drains) [m] 20

Xe (average spacing between emergency exits) [m] 300

Ecom (emergency coms) → 1, 2 o 3 [-] 3

Tunnel Data






Tunnel St. Demetrio: F-N curve in the south direction

1,00E-06

1,00E-05

1,00E-04

1,00E-03

1,00E-02

1,00E-01

1,00E+00

1,00 10,00 100,00 1000,00

F CU

M [a

cc/y

ear]

N [FAT]

Tollerable Risk Line

Acceptable Risk LineEV [Fat/year] 1.68E-02






Tunnel St. Demetrio: QRAM Sensitivity analysis

input parameter Variation Initial Value Final Value Initial Value Final Value

0 0.01 0.00 0.01 0.00

equal to HGV ratio 0.01 0.02; 0.1; 0.117 0.01 0.02; 0.1; 0.117

0.15 - 0.15 0.30 ; 0.00 0.15 ; 0.15 0.30 ; 0.00 0.15 ; 0.15

0 - 0.30 0.30 ; 0.00 0.00 ; 0.30 0.30 ; 0.00 0.00 ; 0.30

1 2 1 2 1

1.5 2 1.5 2 1.5

2.5 2 2.5 2 2.5

3 2 3 2 3

1.5 1.1 1.5 1.1 1.5

2 1.1 2 1.1 2

3 1.1 3 1.1 3

x 10 1.61E-07 1.61E-06 1.60E-07 1.60E-06

x 10-1

1.61E-07 1.61E-08 1.60E-07 1.60E-08

x 10-1

1.00 0.10 1.00 0.10

x 10 1.00 10.00 1.00 10.00

2.5 0.00 2.50 0.00 2.50

4.12 0.00 4.12 0.00 4.12

1 3 1 3 1

2 3 2 3 2A,B: 0 0.32 0.00 0.32 0.00A,B: 3 0.32 3.00 0.32 3.00

A: -0.32 0.32 -0.32 - -

B: -0.32 - - 0.32 -0.32

A,B: -3 0.32 -3.00 0.32 -3.00

1 2 1 2 1

3 2 3 2 3

XI Type of Construction (1 Circular, 2 Rectangualar cross-section) 2 [-] 1 2 1 2

105 0.00 105.00 0.00 -105.00

210 0.00 210.00 0.00 -210.00

200 261.00 200.00 -261.00 -200.00

300 261.00 300.00 -261.00 -300.00

REVERSE 261.00 -261.00 -261.00 261.00

0 0.075 0.00 0.075 0.00

0.15 0.075 0.15 0.075 0.15

1 3 1 3 1

2 3 2 3 2

200 300.00 200.00 300.00 200.00

400 300.00 400.00 300.00 400.00

1 150 1 150 1

2 150 2 150 2

3 150 3 150 3

4 150 4 150 4

5 150 5 150 5

10 150 10 150 10

XVII

Safety

equipment

XII

XIII

XIV

XV

XVI

VII

VIII

IX

X

Changes to

the

structure

Frequency

of

Accidents

I

II

Bus Coaches Ratio (for each period: QUIET, NORMAL, PEAK)

Propane in Bulk ratio - Propane in Cylinder ratio

Average Number of People in a Light Vehicle

Average Number of People in a HGV

III

IV

TRAFFIC

V

VI

Average Spacing between Emergency Exits

Delay for Stopping Approaching Traffic

Accidents Frequency (facc)

DG-HGV transport correction factor

Camber (transversal slope)

Ground Type: 1 (Bedrock), 2 (Fragmented), 3 (Fragmented and

Under Water

Segments Gradient

Number of Lanes

[m3/s]Normal Longitudinal Ventilation, Volume Flow Rate along tunnel (at

each node)

Emergency Longitudinal Ventilation, Volume Flow Rate along tunnel

(at each node)

Open Area of discrete Drains

Emergency Coms: 1 (bell/siren), 2 (Public Address system)

[min]

[-]

[-]

[m2]

[m3/s]

Societal Risk A - SOUTH Syr B - NORTH Cat

[-]

[%]

[-]

[%]

[-]

[-]

[acc/(veh*

km*year)]

[pass]

[pass]

[-]






1,00E-03

1,00E-02

1,00E-01

initia

l cu

rve

Bu

s r

atio =

0

Bu

s C

oach

es R

atio =

HG

V R

atio

LP

G in

Bu

lk =

LP

G in C

ylin

de

r =

0.1

5

LP

G in

Cylin

der

= 0

.30

Pe

ople

in a

Lig

ht V

ehic

le =

1

Pe

ople

in a

Lig

ht V

ehic

le =

1.5

Pe

ople

in a

Lig

ht V

ehic

le =

2.5

Pe

ople

in a

Lig

ht V

ehic

le =

3

Pe

ople

in a

HG

V =

1.5

Pe

ople

in a

HG

V =

2

Pe

ople

in a

HG

V =

3

facc x

10

facc x

10-1

DG

-HG

V corr

ection

facto

r *

10-1

DG

-HG

V tra

nsp

ort

co

rre

ction facto

r *

10

Cam

be

r =

2.5

Ca

mb

er

= 4

.12

Gro

un

d (

Bad

Rock):

1

Gro

un

d T

ype (

Fra

gm

ente

d):

2

Se

gm

en

t G

radie

nt =

0

Se

gm

en

t G

radie

nt =

3

Se

gm

en

t G

radie

nt (

SO

UT

H)

= -

0.3

2

Se

gm

ent G

radie

nt (N

OR

TH

) =

-0

.32

Se

gm

en

t G

radie

nt =

-3

Num

be

r o

f La

ne

s 1

Num

be

r o

f La

nes 3

Constr

uction 2

(R

ecta

ngu

ala

r cro

ss-s

ection)

Norm

al Longitud

ina

l V

en

tila

tio

n 1

05

Norm

al Longitud

ina

l V

en

tila

tio

n 2

10

Ope

n A

rea o

f dis

cre

te D

rain

s =

0

Ope

n A

rea o

f dis

cre

te D

rain

s *

2

Em

erg

ency C

om

s =

1 (

be

ll/sire

n)

Em

erg

ency C

om

s =

2 (

Public

Ad

dre

ss s

yste

m)

Em

erg

ency L

on

gitud

inal V

en

tila

tion 2

00

Em

erg

ency L

on

gitud

inal V

en

tila

tion 3

00

Em

erg

ency L

ongitudin

al V

entila

tion →

Revers

e F

low

Ave

rage S

pacin

g b

etw

ee

n E

merg

ency E

xits =

200

Ave

rage S

pacin

g b

etw

ee

n E

merg

ency E

xits =

400

Dela

y for

Sto

pp

ing T

raffic

= 1

min

Dela

y for

Sto

pp

ing T

raffic

= 2

min

Dela

yfo

r S

top

pin

g T

raffic

= 3

min

De

lay fo

r S

topp

ing T

raff

ic =

4 m

in

Dela

y for

Sto

pp

ing T

raffic

= 5

min

Dela

y for

Sto

pp

ing T

raffic

= 1

0 m

in

EVs in Direction South

Tunnel St. Demetrio: Sensitivity analysis results

Traffic

Frequency

of accidentsStructure

details

Safety

equipment

Number

of Lanes

facc x

10

DG-HGV

factor x 10

Delay for stopping

approaching trafficBUS ratio






1,00E-07

1,00E-06

1,00E-05

1,00E-04

1,00E-03

1,00E-02

1,00E-01

1,00E+00

1,00 10,00 100,00 1000,00

initial curve

Delayfor Stopping Approaching Traffic = 1 min

Delay for Stopping Approaching Traffic = 5 min


Acceptable Risk Line





Sensitivity to parameter: “delay for stopping approaching traffic”







1,00E-07

1,00E-06

1,00E-05

1,00E-04

1,00E-03

1,00E-02

1,00E-01

1,00E+00

1,00 10,00 100,00 1000,00

initial curve

Number of Lanes (for every section and in both directions) = 1

Number of Lanes (for every section and in both directions) = 3



𝑁 = 𝑅 ∙ 𝐷𝑅𝑈𝐽 ∙ %𝐿𝐸𝑇𝐻




Sensitivity to parameter: “number of lanes”



1,00E-07

1,00E-06

1,00E-05

1,00E-04

1,00E-03

1,00E-02

1,00E-01

1,00E+00

1,00 10,00 100,00 1000,00

initial curve

facc x 10

facc x 10-1



𝑓𝑖𝑗 = 𝑃𝑖𝑗 ∙ 𝑓𝑎𝑐𝑐,𝑖 ∙ 𝑇𝐻𝑖 ∙ 𝐿𝑖 ∙ 24 ∙ 365 ∙ 10−6




Sensitivity to parameter: “HGVs accident frequency”



1,00E-07

1,00E-06

1,00E-05

1,00E-04

1,00E-03

1,00E-02

1,00E-01

1,00E+00

1,00 10,00 100,00 1000,00

initial curve

Propane in Bulk ratio =Propane in Cylinder ratio =0.15

Propane in Cylinder ratio = 0.30






Sensitivity to parameter: “LPG (Liquefied Petroleum Gas) ratio”



conclusions

The parameters that most affect the risk curve:

• Density of people on the road

• Traffic (veh/h)

• Bus ratio (%)

• Number of lanes

• Delay for stopping approaching traffic

• Average vehicle occupancy

• Accident scenarios frequency [scen/year]

• facc

• DG-HGV traffic

• HGV traffic

• Proportion of each DG

𝑓𝑖𝑗𝑘 = 𝑃𝑖𝑗𝑘 ∙ 𝑓𝑎𝑐𝑐_𝐷𝐺,𝑖 ∙ 𝑇𝐷𝑖𝑘 ∙ 𝐿𝑖 ∙ 24 ∙ 365 ∙ 10−6

• Further risk mitigation measures (adopted only after a cost benefit analysis)

• The safety margin is high

San Demetrio Tunnel Risk Analysis

General Conclusions on the PIARC/OECD QRA model







CONCLUSIONS: QRAM AND Fluid Dynamics/evacuation models

Data Collection

Data Preparation

Risk Calculation

Using QRAM

Is Risk

acceptable?

NO

Additional

risk reduction

measures

START

YES

End

Idintification of

Critical Scenarios

Single Scenario

Simulation

CFD Simulation

(Fire, Ventilation)

Evacuation Model

(Evacuation, Rescue)

Qualitative Risk

Estimation

Measures

Included

in the model?

YES NO

(Gai et al., Proceedings IF CRASC’ 15)

An operating method to follow can be to

identify the critical scenarios that give the most

significant contribution to the overall risk

through the QRAM, and then to simulate those

scenarios in detail in order to define risk

reduction measures (Petelin S. 2009)







• NFPA, SFPE Handbook of Fire Protection Engineering, 3rd edition, 2002

• Jönsson, J. Combined Qualitative and Quantitative Fire Risk Analysis – Complex Urban Road Tunnel. Arup partners, 2007.

• Faber, M.H. (2008) Risk and Safety in Civil, Environmental and Geomatic Engineering. ETH Zürich, lecture notes, available online on

01/2011 at: http://www.ibk.ethz.ch/fa

• Haimes, Y. Y. (1981). Hierarchical holographic modeling. IEEE Transactions on Systems, Man, and Cybernetics, 11(9), pp. 606– 617.

• Nolan, D.P. 1986. Handbook of Fire and Explosion Protection Engineering Principles for Oil, Gas, Chemical, and Related Facilities.

Noyes, New Jersey

• Aven, T. Risk Analysis: Assessing Uncertainties beyond Expected Values and Probabilities. John Wiley & Sons, 2008

• Furness, A. , Muckett, M. Introduction to Fire Safety Management. Elsevier, 2007.

• Fire Risk in Metro Tunnels and Stations, Hyder Consulting, available on 05.2011 at

http://hkarms.myftp.org/web_resources/Conference_Presentation/Fire_Risk_Metro_Tunnels_Stations.pdf

• Fault Tree Handbook. Nuclear Regulatory Commission. NUREG–0492

• Hasofer et al. 2007, Risk Analysis in Building Fire Safety Engineering

• Persson, M. Quantitative Risk Analysis Procedure for the Fire Evacuation of a Road Tunnel -An Illustrative Example. Lund, 2002

• Brussaard et al. 2004. The Dutch Model for the Quantitative Risk Analysis of Road Tunnels. Available on 05.2011 at

http://www.rws.nl/rws/bwd/home/Tunnelveiligheid/dutch%20model.pdf

• Gkoumas, K. 2008. Basic aspects of risk-analysis for civil engineering structures. Handling Exceptions in Structural Engineering:

Robustezza Strutturale, Scenari Accidentali, Complessità di Progetto, Roma, 13-14 novembre.

http://www.francobontempi.org/handling_papers.php

References (1/2)





http://www.ibk.ethz.ch/fa

http://hkarms.myftp.org/web_resources/Conference_Presentation/Fire_Risk_Metro_Tunnels_Stations.pdf

http://www.rws.nl/rws/bwd/home/Tunnelveiligheid/dutch model.pdf

http://www.francobontempi.org/handling_papers.php



References (2/2)

83

• Di Santo, C., Gkoumas, K., Bontempi. F. “Risk analysis for severe traffic accidents in road tunnels (Part I)”, Ingegneria Forense, Crolli,

Affidabilità Strutturale e Consolidamento. Atti del Convegno IF CRASC '15 - 14/16 maggio 2015, Roma, Nicola Augenti & Franco

Bontempi (ed), Dario Flaccovio Editore, pp. 971-979, ISBN 9788857904474

• Di Santo, C., Gkoumas, K. “Risk analysis for severe traffic accidents in road tunnels (Part II)”, Ingegneria Forense, Crolli, Affidabilità

Strutturale e Consolidamento. Atti del Convegno IF CRASC '15 - 14/16 maggio 2015, Roma, Nicola Augenti & Franco Bontempi (ed),

Dario Flaccovio Editore, pp. 959-969, ISBN 9788857904474

• DG MOVE (2014). Harmonised Risk Acceptance Criteria for Transport of Dangerous Goods. London: Det Norske Veritas Ltd.

• Diernhofer, F., Kohl, B. and Hörhan, R. (2010). New Austrian Guideline for the Transport of Dangerous Goods through Road Tunnels. 5th

International Conference on Tunnel Safety and Ventilation, Graz, 2010.

• Evans, A.W. (2003). Transport Fatal Accidents and FN-curves: 1967-2001. HSE Research Report 073. HSE Books: Sudbury,

Suffolk.Directive 2004/54/EC of the European Parlament and the Concil of 29 April 2004 on minimum safety requirements for tunnels in

the trans-European road network, 7.6.2004

• Impresa Pizzarotti & C. S.p.A. (2007). Affidamento dell’opera di completamento del tratto stradale Catania Siracusa con caratteristiche

autostradali, compreso tra le località passo s. martino ed Il Km 130+400 Della Ss. 114, Analisi Di Rischio (in Italian).

• Melchers, R.E. 1993. Society, tolerable risk and the ALARP principle. In: Melchers RE, Stewart MG, editors. Probabilistic risk and

hazard assessment. Netherlands, Balkema: 243–252

• Petelin S., Luin, B. and Vidmar, P. (2010). Risk Analysis Methodology for Road Tunnels and Alternative Routes. Journal of Mechanical

Engineering, Vol. 56, No. 1, pp. 41-51.

• PIARC Technical Committee C3.3 (2008) Road Tunnel Operation, Risk Analysis for Road Tunnels, ISBN 2-84060-202-4.

• PIARC-OECD (2001). Safety in Tunnels, Transport of dangerous goods through road tunnels. OECD Publications.

• UNCED AGENDA 21. 1992. Program of Action for Sustainable Development. United Nations Publication - Sales No. E.93.I.

• Vagiokas, N., Bletsas, A. and Nelisse, R.M.L. (2013) Methodological approaches for tunnel classification according to ADR agreement.

World Tunnel Congress 2013 Geneva Underground – the way to the future! G. Anagnostou & H. Ehrbar (eds).



