gas as a marine fuel...lng during bunkering. it specifically looks at how the safety zone can be...
TRANSCRIPT
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
gas as a marine fuel Recommendation of Controlled Zones during LNG bunkering.
safetyversion 1.0 FP02-01
© Society for Gas as a Marine Fuel
Version 1.0, May 2018.
© Society for Gas as a Marine Fuel, 2018.
ISBN: 978-0-9933164-8-7
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the Society for Gas as a Marine Fuel.
Disclaimer
While the advice given in this “Recommendation of Controlled Zones during LNG bunkering” has been developed using the best currently available information, it is intended solely as guidance to be used at the owner’s own risk.
Acknowledgements
This document was produced by SGMF’s Working Group 2. SGMF acknowledges the participation of the following individuals and companies in its development:Sean Bond (American Bureau of Shipping), Cees Boon (Port of Rotterdam), John Boreman (BP Shipping), Dimitrios Dalkakis (World Maritime University), Charles de Souza (MPA Singapore), Guenter Eiermann (Nauticor), John Eltringham (Bernhard Schulte Shipmanagement), David Haynes (SGMF), Daniel Jaerschel (Dubai Supply Authority), Mike Johnson (DNV GL), Marcel LaRoche (British Colombia Ferries), Eric Linsner (Marshall Islands Registry), Roland Peeters (previously with Falck), Gema Lopez Garrido (Gas Natural Fenosa).
SGMF would also like to acknowledge the contributions of the following individuals and organisations:Franklin Salazar and Gagan Lamba (both BP).
© Society for Gas as a Marine Fuel I
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
The LNG industry has an excellent safety record. This has been achieved because the industry understands the hazards associated with LNG and how these can be managed effectively. As the maritime industry moves towards using LNG as a fuel, it is important that this knowledge is also transferred. In this respect, guidance developed by SGMF plays a key role.
One critical difference between LNG and other maritime fuels is that in the highly unlikely event of an accidental spillage, the hazard can extend some distance from the LNG installation because of the formation of a gas cloud. Controlling the hazard requires an understanding of the extent of the zone that might be affected and the measures that can be taken to reduce risks.
Previously published SGMF guidance for LNG bunkering operations defines a number of zones, most of which do not need quantification in terms of distance. However, we have thought long and hard about how to define the “Safety Zone” – which does need to be quantified.
We understood what we needed: it had to be simple to apply, representative of the hazard, and practicable for the industry. Importantly, it also had to be based on a methodology understood by SGMF. We concluded that explicit quantified risk assessment was not the route to take because, generally, it is not simple to apply, there are significant uncertainties regarding the likelihood of any release, and it is difficult to define a tolerable level of risk.
Instead, we have based the calculated distances on the consequences of an LNG release, particularly vapour dispersion, where there is much less uncertainty. We used single representative release sizes, determined using the best available information combined with engineering judgement. We deliberately use the term “representative” – avoiding the terms “credible” or “maximum credible”, the meanings of which are problematic.
This has resulted in the development of a safety distance calculator that can be used by SGMF members for many situations, with the option for more detailed analysis for more complex or sensitive operations,
Foreword
© Society for Gas as a Marine Fuel II
if required. We believe that, when combined with existing SGMF guidance, it is a significant step forward in promoting safe LNG bunkering operations.
Mike JohnsonDNV GL / Chairman, SGMF WG2 Recommendation of Controlled Zones during LNG bunkering
© Society for Gas as a Marine Fuel III
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
This safety document provides guidance on how to determine the size and location of controlled zones around the bunkering infrastructure of an LNG supplier and gas-receiving ship to facilitate the safe transfer of LNG during bunkering. It specifically looks at how the Safety Zone can be calculated and implemented. It covers only bunkering, the transfer of LNG to a gas-fuelled vessel, and, where relevant, the handling of vapour return.
LNG and natural gas behave differently from traditional fuel oils when released into the air or onto water or land. So safety precautions have to be assessed differently than for traditional bunkering operations. The guidelines address the following operational scenarios:
• ship-to-ship bunkering
• truck-to-ship bunkering
• shore-based terminal-to-ship bunkering
Summary
© Society for Gas as a Marine Fuel IV
Controlled Zones
Five controlled zones are defined below and shown in the Figure.
LNG BUNKERING ZONES ILLUSTRATION(Truck to ship Method shown as example)
III MONITORING AND SECURITY AREA
I HAZARDOUS ZONE
II SAFETY ZONE
IV MARINE ZONE
V EXTERNAL ZONE
III
I
II
1: Hazardous ZoneThe Hazardous Zone is a three-dimensional space in which a combustible or explosive atmosphere can be expected to be present frequently enough to require special precautions for the control of potential ignition sources. Hazardous zones are always present but addressed via appropriate design techniques and safety practices.
Hazardous zones must be defined for all components of the LNG bunkering system by their respective owners. These components can include gas-fuelled ships, bunker vessels, road tankers and terminals.
The LNG supplier, or the infrastructure owner on the supplier’s behalf, is responsible for checking and confirming the compatibility of the multiple
© Society for Gas as a Marine Fuel V
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Hazardous Zones between the infrastructure and the gas-fuelled ship. This check must be performed for each new combination of gas-fuelled vessel, location and bunkering infrastructure configuration. If individual zones do not add up to make an acceptable combined zone, bunkering should not proceed.
2: Safety ZoneThe Safety Zone can be defined as the three-dimensional envelope of distances inside which the majority of leak events occur and where, in exceptional circumstances, there is a recognised potential for a leak of natural gas or LNG to harm life or damage equipment/infrastructure.
The zone is temporary by nature, present only during bunkering. It may extend beyond the gas-fuelled ship/LNG road tanker/bunker vessel, interconnecting pipework, and so on, and will be larger than the Hazardous Zone.
The purpose of the Safety Zone is to minimise the likelihood of harm to people and damage to equipment by:
• controlling leaks and spills
• avoiding ignition and a subsequent fire or explosion
• excluding non-essential people (to avoid additional injuries or deaths in the event of an accident)
• protecting essential staff through the use of PPE (to minimise the likelihood of injury or death in the event of an accident)
The Safety Zone should always be under the control of the Person In Charge (PIC). It must therefore lie within the port or another entity that allows the PIC the required degree of control.
The size of the Safety Zone will depend on:
• the design of the LNG bunkering infrastructure/gas-fuelled ship
• the configuration of the LNG transfer system
• the duration, flow-rate and pressure of the potential leak source
© Society for Gas as a Marine Fuel VI
• weather conditions and ambient temperature
• the layout of the location where spills could occur
3: Marine Exclusion ZoneThe purpose of the Marine Exclusion Zone is to protect the bunkering vessel from other marine traffic, primarily by defining minimum distances and speeds for passing vessels.
Definition of the Marine Exclusion Zone is for each port to decide and implement in port rules, based on specific port and ship studies. All ships and bunker vessels must comply with these rules in the normal way.
4: Monitoring & Security AreaThe Monitoring & Security Area is defined as the three-dimensional space inside which activities (including people and vehicle movements) need to be identified and monitored to ensure that they do not affect the safety of the bunkering operation by encroaching on the Safety Zone of the gas-fuelled ship, quayside or LNG bunkering infrastructure. Its primary purpose is to prevent impacts from the actions of people not involved in the bunkering process.
The Monitoring & Security Area will always be larger than the Safety Zone. As the reasons for the Monitoring & Security Area are many and wide-ranging, it is unlikely that it will be possible to define or justify the size of the Monitoring & Security Area by calculation. It should be considered as a contingency on, or factor to, the Safety Zone. This area is only relevant during bunkering.
5: External ZoneIn some jurisdictions – for example, much of Europe – an External Zone is required. A port cannot influence how the general public behaves outside the port area so the risk level outside must be kept low. This zone is defined by the level of risk general members of the public can be exposed to, based on local regulatory requirements.
© Society for Gas as a Marine Fuel VII
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Hose, Piping, Valve and Flange Leaks
SGMF has conducted research to develop a calculation methodology that allows the Safety Zone dimensions to be conservatively estimated – reliably and consistently – for a wide range of bunkering configurations, flow rates and locations.
After extensive consultation with SGMF members and cryogenic industry bodies, SGMF has yet to identify a failure mode where a hose or loading arm rupture, where the failure creates a hole so large that it covers most of the surface area of the hose or breaks the hose in two (also called a guillotine failure) is likely to occur. All the evidence from the industry suggests that hoses fail before rupture, allowing the transfer to be stopped and the hose removed from service and disposed of. The only possibility of a hose or loading arm rupture is from a vessel or road tanker pulling away with the transfer system still attached. The IGF Code (Section 8.4) has a requirement for a breakaway device that will split the loading arm/hose in a controlled manner and keep LNG outflow to a minimum. In the light of this requirement, a hose rupture is not considered in these guidelines.
SGMF has collated a variety of experience (on a confidential basis) from the filling of LNG road tankers and cryogenic gas (liquid nitrogen, liquid oxygen) operations to assess good practice. About five million LNG truck transfers are estimated to have occurred. The findings of this work are that:
• no guillotine failures have been recorded
• hose handling determines hose life; the main cause of failure is fatigue from temperature, pressure and movement cycles, particularly bending
• fatigue failures produce small leaks – at least initially – so they are detectable at an early stage, allowing the hose to be removed from service and destroyed
• small leaks can occur through poor hose management, such as inadequate leak checking, worn couplings damaged gaskets, dirt and debris on flange faces and gaskets, and operator errors and bad habits
© Society for Gas as a Marine Fuel VIII
There is no reliable information available describing how LNG hoses fail. Using limited experimental work, information from other industries, and reliability assessment work on piping in the offshore oil and gas industry and failure behaviour in rubber LPG/oil hoses, a link between hole size to pipe/hose size has been developed. Given the available information, SGMF estimates that a hole size of 6% of the hose diameter is appropriate. Anecdotal comments from industry experience suggest this is conservative.
If the design of the hose can be demonstrated to give a significant reduction in the likelihood of releases from the hose itself, then releases from the flanges and valves at either end of the hose will dictate the release distance. In this case, the Safety Zone should be based on releases from these locations.
SGMF recommends that the following hole sizes are appropriate for estimating the size of the LNG/gas cloud/pool and therefore the size of the Safety Zone.
Hole sizes used in SGMF’s Bunkering Area Safety Information for LNG (BASiL) model
Size Metal or composite hose
Fitting/gasket/valve, fixed piping or hard arm
2 inch / 50 mm 3 mm 3 mm
3 inch / 75 mm 4.5 mm 4 mm
4 inch / 100 mm 6 mm 4 mm
6 inch / 150 mm 9 mm 4 mm
8 inch / 200 mm 12 mm 5 mm
10 inch / 250 mm 15 mm 5 mm
© Society for Gas as a Marine Fuel IX
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
These hole sizes are based on the following criteria:
• the Safety Zone is under the control of the PIC of the bunkering operation; when the PIC or another operative identifies a leak, the system will be shut down to limit the consequences of the incident
• management systems have been developed and implemented to ensure that damaged hoses are identified prior to use through inspection
• if a leak is found during inspection or during the bunkering operation, and cannot immediately be sealed, the hose should be immediately removed and destroyed
• the scenario of a ship and bunkering facility (vessel or road tanker) moving away from each other is controlled both by the procedures and requirements in the IGF Code for a dry-break coupling
• the potential for other impact damage is controlled by appropriate procedures, particularly in relation to allowable operations being conducted simultaneously with LNG bunkering (SIMOPs)
Leak Behaviour Gas clouds formed by leaking LNG can travel significant distances before they ignite. The Safety Zone is defined by the maximum distance the gas evaporating from a pool of LNG or from a pressurised LNG release can subsequently be ignited, based on the hole sizes above. On this basis, the delayed ignition of a gas cloud causing a flash fire is argued to be the event that defines the safety distance.
Many factors determine how far a gas cloud will spread and remain within flammable limits. The parameters considered by SGMF include:
• LNG transfer flow rate, temperature and pressure
• hole size
• different orientations of leaks – vertically, horizontally and downwards
© Society for Gas as a Marine Fuel X
• various climatic conditions around wind speed, climatic stability, ambient temperature and humidity
• a range of LNG compositions and physical properties
• different geometries/topographies for releases over land and water and at different elevations
• various durations of release (depending on the type of emergency shut-down system)
Importantly, these parameters interact with each other. This means that some effects must be considered together, which increases the complexity of the model and results in the need to consider some 1.4 million data points for eight parameters.
Safety Zone Distance Calculations SGMF has created a model called Bunkering Area Safety Information for LNG (BASiL) to estimate the size of the Safety Zone based on the extent of the gas cloud to 100% LFL. The BASiL model is available on SGMF’s website (www.sgmf.info).
BASiL estimates several distances to LFL which, when combined together in the model, produce a three-dimensional envelope defining the Safety Zone.
Reviews of accuracy against the dispersion models have shown that the vast majority of the BASiL calculations result in safety distances within ±10% of the actual values.
© Society for Gas as a Marine Fuel XI
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Contents
Foreword .................................................................................I
Summary ............................................................................... III Controlled Zones ........................................................................IV Hose, Piping, Valve and Flange Leaks ....................................VII Leak Behaviour ...........................................................................IX Safety Zone Distance Calculations ........................................... X
Abbreviations and Definitions .............................................XIII
1. Purpose and Scope .................................................... 1
2. LNG Hazards ............................................................. 3
3. Identifying and Understanding Bunkering Risks ...... 4
4. Controlled Zones ....................................................... 6 4.1. Hazardous Zone .................................................................6 4.2. Safety Zone ......................................................................... 11 4.3. Marine Exclusion Zone ..................................................... 13 4.4. Monitoring & Security Area .............................................. 13 4.5. External Zone .................................................................... 15 4.6. Roles of Bunkering Stakeholders in Defining Controlled Zones ....................................................................... 16
5. Calculating Distances for the Safety Zone ............. 20 5.1. Scenario Selection ............................................................. 20 5.2. Failure Scenarios .............................................................. 22 5.3. Leak Behaviour ................................................................. 23 5.4. The SGMF Safety Distance Model – BASiL ...................... 29 5.5. Example Calculations ...................................................... 36
6. Calculating Distances for the Monitoring & Security Area ........................................................... 42
7. SIMOPS and Non-Standard Operations ................. 44
© Society for Gas as a Marine Fuel XII
8. Bibliography ............................................................ 46 8.1. Publications ....................................................................... 46 8.2. Relevant International Standards .................................... 46
Appendix A: Hazards Associated With LNG ....................... 48 A.1. LNG ................................................................................... 48 A.2. Boil-Off Gas and Natural Gas ......................................... 49 A.3. LNG Leak Sources ............................................................ 49 A.4. LNG Leak Behaviour ......................................................... 51 A.5. Ignition Scenarios ............................................................ 53 A.6. LNG Fires ........................................................................... 56 A.7. Explosions ......................................................................... 58 A.8. Other Hazard Scenarios .................................................. 59
Appendix B: Hazard Identification ...................................... 60
Appendix C: Calculating Distances for the Safety Zone ..... 74 C.1. Scenario Selection ............................................................74 C.2. Leak Behaviour ..................................................................76
Appendix D: Hole Sizes in LNG Transfer Systems ................ 91 D.1. Failure of Cryogenic Hoses ............................................. 95 D.2. Hole Size Selection ........................................................... 99
Appendix E: BASiL Validation ............................................. 107
© Society for Gas as a Marine Fuel XIII
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Abbreviations and Definitions
ADN – Accord Européen Relatif au Transport International des Marchandises Dangereuses par Voies de Navigation Intérieures. (European Agreement Concerning the International Carriage of Dangerous Goods by Inland Waterways)
ADR – Accord Européen Relatif au Transport International des Marchandises Dangereuses par Route. (European Agreement Concerning the International Carriage of Dangerous Goods by Road)
ALARP/ALARA – As Low As Reasonably Practicable/Achievable without incurring excessive costs
Alkane – A member of the family of saturated hydrocarbon compounds consisting solely of carbon and hydrogen atoms with all bonds between them consisting of single bonds. The most common are methane, ethane, propane and butane
ATEX – Appareils destinés à être utilisés en ATmosphères EXplosibles. (Equipment destined for use in potentially explosive atmospheres.) The European Union ATEX Directive 2014/34/EU covers equipment and
protective systems intended for use in potentially explosive atmospheres
Atmospheric stability – A measure of the atmosphere’s tendency to encourage or deter vertical motion
Auto-ignition temperature – The minimum temperature required to ignite a gas or vapour in air in the absence of a spark or flame
bara – Pressure stated as absolute pressure, meaning the pressure zero-referenced to that of a perfect vacuum
barg – Pressure stated as gauge pressure, meaning the pressure zero-referenced to ambient atmospheric pressure
BAT/BACT – Best Available Technology/ Best Available Control Technology. The environmental equivalent of ALARP
BOG – Boil-Off Gas. The vapour created by evaporation from the surface of a volume of LNG
CCNR – Central Commission for Navigation of the Rhine. The body that controls regulations on the major international inland waterways of Europe
© Society for Gas as a Marine Fuel XIV
Competent Authority – Any national, regional or local authority or other authorities empowered, alone or together, to act as the regulatory body on LNG bunkering
Controlled Zones – Areas extending from the bunkering equipment, pipework connections, and emergency vent locations that have restrictions in place continuously or during bunkering
Dry Breakaway Coupling – A safety coupling located in the LNG transfer system which separates at a predetermined break-load. Each separated section contains a self-closing shut-off valve, which seals automatically to prevent any spill during a breakaway. Also known as a dry disconnect coupling
Dry Disconnect Coupling – See Dry Break-away Coupling above
Emergency Release Coupling (ERC) – A coupling installed on LNG and vapour lines, as a component of the Emergency Release System (ERS), to ensure the quick physical disconnection of the transfer system from the unit to which it is connected. It is designed to prevent damage to loading/unloading
equipment in the event that the transfer system’s operational envelope and/or parameters are exceeded beyond a predetermined point
Emergency Relief System – A system that relieves the pressure within a pipe or storage tank by allowing a fluid to be transferred to another location, normally the atmosphere, when the pressure exceeds a set limit. Relief to atmosphere is only allowed under emergency scenarios where equipment may be damaged
EN – European (Standard) Norm
ESD – Emergency Shut-Down. A control system and associated components that when activated stop operations in a controlled manner and return the system to a safe state.
An ESD system may have several sequential stages, with the operation of each stage dependent on the potential consequences of the situation. During bunkering these stages are commonly designated ESD-1 and ESD-2:
• ESD-1 – where transfer of LNG to the bunkering vessel is stopped
© Society for Gas as a Marine Fuel XV
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
• ESD-2 – where the transfer system is disconnected from the bunkering ship
In some ship types there may be additional definitions of the ESD system but these are outside the scope of this document
External Zone – The distance to a defined risk level, frequently places where the public may be present, required by some regulatory regimes
Flag state – The organisation that enforces international regulations, including those relating to safety and pollution prevention, over commercial vessels registered under its flag
Flammable range – The range of hydrocarbon gas concentrations in air between the Lower and Upper Flammable Limits. Mixtures within this range are capable of being ignited and burnt
Flash point – The lowest temperature at which a liquid gives off sufficient vapour to form a flammable mixture with air above the liquid surface
GIIGNL – Groupe International des Importateurs de Gaz Naturel
Liquéfié. The industry group made up of the world’s main LNG importers
Hazardous Zone – The three-dimensional space where there is a probability that a flammable atmosphere is present. Defined by national regulation and both the IGF and IGC codes
HAZID – HAZard IDentification. There are a number of recognised methods for the formal identification of hazards. For example, a brainstorming exercise using checklists where the potential hazards in an operation are identified and gathered in a risk register to be addressed and managed
IACS – The International Association of Classification Societies
IAPH – The International Association of Ports and Harbours
IGC Code – The IMO International Code for the Construction and Equipment of Ships Carrying Liquefied Gases in Bulk
IGF Code – The International Code of Safety for Ships using Gases or other Low-Flashpoint Fuels
© Society for Gas as a Marine Fuel XVI
IMDG – The IMO International Maritime Dangerous Goods Code
IMO – The International Maritime Organization. The United Nation’s maritime regulatory body
ISM – The International Safety Management Code published by the IMO
ISPS – The International Ship and Port Security Code
ISO – The International Organization for Standardisation. An international standard - setting body composed of representatives from various national standards organisations
ISO Container – A container manufactured according to specifications from ISO which define its size, strength, and durability requirements
LFL/LEL – Lower Flammable Limit/Lower Explosive Limit. The low end of the concentration range over which a flammable mixture of gas and vapour in air can ignite at a given temperature and pressure (see also UFL/UEL)
LNG – Liquefied Natural Gas. Natural gas that has been cooled to the point where it is liquid at stated pressure. GNL in French, Spanish and Italian (French Gaz Naturel Liquéfié)
Marine Exclusion Zone – A zone of sufficient size to prevent passing shipping from impacting the LNG transfer operation
Monitoring & Security Area – An area around the LNG transfer equipment that needs to be monitored as a precautionary measure to prevent interference with the LNG transfer operation
Natural gas (NG) – A mixture of hydrocarbon gases, mostly methane, used as a fuel or chemical feedstock. Also used to refer to regasified LNG
NFPA – The National Fire Protection Association. A US-based standards body for fire, electrical and related hazards
NGO – Non-Governmental Organisation. A not-for-profit organisation independent of governments or international governmental organisations
OCIMF – The Oil Companies
© Society for Gas as a Marine Fuel XVII
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
International Marine Forum. An association representing operators of oil tankers and terminals, dealing with safety and environmental issues and specifically associated with mooring and berthing guidelines
OPITO – Offshore Petroleum Industry Training Organisation. The offshore oil and gas industry’s focal point for skills, training and workforce development
PIC – Person In Charge. The person responsible for the management of an operation such as bunkering. There may be several PICs, each responsible for an operation
POAC – Person in Overall Control. The person responsible for the management of the LNG bunkering process and any SIMOPs being undertaken through one or more PICs
Port authority – a governmental, regional or local, usually public body that develops and manages port safety, port infrastructure and other transportation related infrastructure
Port/terminal owner – a company which owns a terminal or port in a wider port area
Port/terminal operator – a company which is operating a terminal within a wider port area
PPE – Personal Protective Equipment
QRA – Quantitative Risk Assessment. A formalised, numerical risk assessment method for calculating a risk level for comparison with defined risk criteria
Risk – A combination of the likelihood of an event and the consequences if the event occurs
Safety Zone – The three-dimensional envelope of distances inside which the majority of leak events occur and where, in exceptional circumstances, there is a recognised potential to harm life or damage equipment/infrastructure in the event of a leak of gas and/or LNG
SGMF – The Society for Gas as a Marine Fuel. An association for companies involved in the use of LNG as a marine fuel
© Society for Gas as a Marine Fuel XVIII
SIGTTO – The Society of International Gas Tanker and Terminal Operators. An organisation representing operators of gas tankers and import and export terminals, covering all liquefied gases in bulk
SIMOP – SIMultaneous OPeration. Defined in this document as “LNG bunkering plus one, or more, other independent operations conducted together within the control of the PIC(s), where the operations may impact, or increase the impacts on personnel safety, ship integrity and/or the environment”
SMS – Safety Management System, as defined by the ISM Code
Surface roughness – The size of obstacles that cause or increase turbulence along the bottom surface of a dispersing gas plume which affect the degree of mixing between the gas and the surrounding air
Thermal relief valves (TRV) – Used to relieve pressure caused by thermal expansion of process fluids in vessels and long lengths of pipework
UFL/UEL – Upper Flammable Limit/Upper Explosive Limit. The high end of the concentration range over which a flammable mixture of gas and vapour in air can ignite at a given temperature and pressure (see also LFL/LEL)
Vapour return line – A connection between the bunkering facility and the receiving ship that allows excess vapour generated during the bunkering operation to be returned to the bunkering facility, removing any need to vent to atmosphere. It is used to control the pressure in the receiving tank due to the liquid transfer, flash gas and boil-off gas generation
© Society for Gas as a Marine Fuel 1
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
This document provides guidance on how to determine the size and location of controlled zones, particularly the Safety Zone, around the bunkering infrastructure of an LNG supplier and a gas-receiving ship to facilitate the safe transfer of LNG during bunkering operations. This safety guide considers only bunkering, the transfer of LNG to a gas-fuelled vessel and, where relevant, the handling of vapour return. It does not cover the use of LNG or vaporised natural gas on board the gas-fuelled vessel during normal operations.
Moreover, this document is specifically about LNG. Many of the comments are equally appropriate to other fuels allowed by the IGF Code. However, each fuel has its own specialities – for example, the cryogenic nature of LNG or the toxic hazard associated with methanol – and therefore this guidance should only be used in its entirety for LNG.
At this early stage of development in the LNG bunkering industry, risks cannot be directly compared with long-established conventional bunkering. So risk assessment and mitigation needs to have a much higher profile. If the industry continues to grow safely and successfully, these additional practices may have less emphasis in future years.
LNG and natural gas behave differently from traditional fuel oils when released into the air or onto water or land. This means that safety precautions have to be assessed differently than for traditional bunkering operations. This guide provides an overview of how the controlled zones around LNG bunkering operations can be defined. It specifically looks at how the Safety Zone can be calculated and implemented.
This publication is a technical book which primarily provides the necessary information for individuals and organisations to start developing operational and safety guidance. The book does not provide rules or definitive safety distances but the framework to base more detailed rules and procedures on.
Parts of these guidelines talk about mitigation for LNG/gas hazards. This does not include emergency response (including fire fighting). How the emergency services approach a LNG/gas scenario is independent of the
1. Purpose and Scope
© Society for Gas as a Marine Fuel 2
precautionary safety distances presented here. Once a hazardous event has occurred, the risk assessments discussed here become irrelevant and emergency service protocols take over. Guidance for the emergency services is being developed or has been covered elsewhere by other industry bodies (such as SIGTTO, CCNR and OPITO).
These guidelines address the following operational scenarios:
• ship-to-ship bunkering
• truck-to-ship bunkering
• shore-based terminal-to-ship bunkering
More details of each are provided in SGMF’s publication “FR07-1 gas as a marine fuel, safety guidelines, bunkering”. Portable LNG tanks – such as ISO containers used as fuel tanks – are outside the scope of this guidance.
© Society for Gas as a Marine Fuel 3
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
This guide is about defining safety distances and other zones. An explanation of the calculations underlying these Safety Zones requires an understanding of some of the physical phenomena associated with the use of LNG as a fuel.
LNG: • is a boiling liquid which, because of its low storage temperatures
(-140 to -160°C, depending on pressure), is continuously vaporising into gas (boil-off gas)
• will vaporise and rapidly pressurise a system to bursting point if left trapped between two valves without pressure relief
• will damage ship quality steels in the area immediately around a spill; rapid cooling reduces the ductility of steel and its ability to support load, which can cause brittle fracture of a vessel’s deck or of a steel component of a quayside
• may cause a Rapid Phase Transition (RPT) if it hits a water surface and boils so rapidly that an over-pressure situation occurs; an RPT is effectively a flameless explosion of limited power
The boil-off gas (including vapour return):• is heavier than air until it warms to -110°C; so, if it leaks, it will initially
flow downwards
• is flammable at concentrations between 5% and 15% in air
• is not odourised and so does not smell like pipeline natural gas
• is not itself visible but does cause the surrounding water vapour to condense, which produces a visible white cloud
• can lead to a lack of visibility within a vapour cloud
• can be cold enough to cause hypothermia, cold burns and frostbite
• may replace oxygen within a vapour cloud preventing people from breathing (asphyxiation)
More detailed descriptions of LNG and natural gas behaviours, covering a wider range of scenarios and encompassing calculations for all the zones, are provided in Appendix A.
2. LNG Hazards
© Society for Gas as a Marine Fuel 4
Transferring LNG from one container – for example a tank – to another creates the possibility of a leak which could injure humans and damage the local environment. The risks associated with this transfer will depend on how and where the transfer takes place.
Risk assessment takes place at different levels, depending on the maturity, or otherwise, of the bunkering operation. The owner/operator of the transfer location must be involved in the risk assessment during any planning phase/compatibility assessment but may only require a “right to object” to risk assessments performed on existing bunkering plans as a result of minor technical changes. A risk assessment should be made prior to each bunkering to confirm that transfer and local conditions remain controlled and within statutory requirements.
A more extensive risk assessment will be required each time a gas-fuelled vessel:
• bunkers in a new location
• uses a different transfer method – for example, from a bunker vessel when previously it has been fuelled by road tankers
• changes one of the components of the bunkering system
• requests different SIMOPs
or when a port area:
• changes its operations or the size/frequency of passing vessels
• requests different SIMOPs
ISO 18683 provides guidance on performing bunkering operation risk assessments. The use of suitably qualified and/or experienced individuals or consultants to support the risk assessment process, particularly for first applications by an organisation, should be considered.
3. Identifying and Understanding Bunkering Risks
© Society for Gas as a Marine Fuel 5
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Once initial risk assessments have been completed, much of the risk assessment can be standardised to reduce time and complexity. However, there are inherent dangers in this approach where changes to one part of a risk assessment influence another and this is not recognised by participants. If carefully reviewed and thought out, risk assessment simplifications can be made – for example, the choice of hose size and its associated equipment primarily defines the flow rates and pressures concerned and therefore the size of the controlled zones.
The guidance in this document is intended to give consistent management of the risks. A range of transfer leak scenarios and their consequences have been examined and the risks assessed. Expert opinions have been taken into account regarding the likelihood and scale of LNG releases and their potential impact on people and equipment.
The resulting guidance – automated through SGMF’s new Bunkering Area Safety Information for LNG (BASiL) model – can be used to manage bunkering risks on a consistent basis, through the definition of a safety zone that depends on the type of bunkering operation being undertaken. Unless unusual or specific circumstances beyond the scope of this document are being considered, this guidance can be used as an alternative to an explicit, detailed risk assessment.
© Society for Gas as a Marine Fuel 6
4. Controlled Zones
This chapter examines how risk assessments can be translated into exclusion areas/zones and how these should be defined to ensure that the risks involved in LNG bunkering are minimised.
Five exclusion distances/zones are defined:
1. Hazardous Zone The zone prescribed by regulation where a flammable atmosphere may be present at any time.
2. Safety Zone The zone, extending beyond the Hazardous Zone, where special precautions are required because of the hazards presented by natural gas/LNG during bunkering operations.
3. Marine Exclusion Zone The zone where passing ships may affect the safety of the bunkering operation.
4. Monitoring & Security Area The area where activities, including shore-side traffic, should be monitored to ensure that they do not encroach on the Safety Zone.
5. External Zone The zone where the risks to individuals, particularly the general public, are controlled by local regulations.
These zones are described in more detail in the following subsections.
4.1. Hazardous Zone
The Hazardous Zone is defined as:The three-dimensional space in which a combustible or explosive atmosphere can be expected to be present frequently enough to require special precautions for the control of potential ignition sources
© Society for Gas as a Marine Fuel 7
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
For LNG bunkering, combustible gas means any vapour leaking from the boiling LNG. This vapour, effectively natural gas, is only explosive in a confined space – for example, where a bunkering manifold is located within a compartment of a vessel. Methane, the main constituent of LNG/natural gas, is flammable only in the concentration range of 5% to 15% in air.
Ignition sources are normally assumed to be electrical equipment (including mobile/cell phones and high-power radio and radar) but can also include:
• static electricity generated by the pumping of liquids or the loading of cargoes using conveyor belts
• naked flames from welding, paint stripping and people smoking
• vehicles (particularly gasoline/petrol-fuelled) delivering to a vessel or boarding a ferry
Examples of special precautions required to limit the probability of ignition sources coming into the hazardous area include:
• using intrinsically safe equipment, which cannot spark
• prohibiting people from bringing ignition sources within the area
Area classification has been developed to categorise explosive gas atmospheres, ensuring the correct selection and installation of equipment known to operate safely in a particular environment. Categories are sub-divided to take into account the properties of the flammable materials present. The pressure of the gas is important in determining the size of the Hazardous Zones. The zone definitions take no account of the consequences of a release. In Europe, ATEX extends the analysis to take into account non-electrical sources of ignition and mobile equipment that creates an ignition risk.
© Society for Gas as a Marine Fuel 8
Hazardous areas are categorised into the following zones:
Table 4.1 Hazardous area classification
EventEuropean, IGC and IGF Code
Definitions
US Definition (NFPA 70)
Time guidance(not officially adopted)
An area in which an explosive gas atmosphere is present continuously or for long periods
Zone 0 Class 1Division 1
Explosive atmosphere for more than 1,000 hours per year
An area in which an explosive gas atmosphere is likely to occur in normal operation
Zone 1 Class 1Division 1
Explosive atmosphere for more than 10, but less than 1,000 hours per year
An area in which an explosive gas atmosphere is not likely to occur in normal operation and, if it occurs, will only exist for a short time
Zone 2 Class 1Division 2
Explosive atmosphere for less than 10 hours per year, but still sufficiently likely to require controls over ignition sources
An area in which an explosive gas atmosphere will not occur in normal operation
Non-hazard-ous or Safe
area
No explosive atmosphere present
The classification of electrical equipment needs additional data based on the combustion properties of the gas involved.
Table 4.2: Gas groups for hazardous area classification
Location Gas group Temperature designation
Europe/International IIA (methane)T1 (above 450°C)
North America D (methane)
Where several hazardous areas exist – for example, a road tanker and loading hose coupled to a gas-fuelled ship – they must not be considered
© Society for Gas as a Marine Fuel 9
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
in isolation. In this example, the two Hazardous Zones must be combined to produce a single Hazardous Zone. This Hazardous Zone exists in three dimensions and may extend beyond the gas-fuelled ship/LNG road tanker/bunker vessel, interconnecting pipework, ISO containers, and so on. Figure 4.1 shows a hazardous area drawing for a LNG bunkering vessel.
Figure 4.1: Hazardous area drawing for a 7,500 m3 bunkering vessel© Bernhard Schulte Shipmanagement
KEYHazardous Area Zone 0
Hazardous Area Zone 1
Hazardous Area Zone 2
Non hazardous
The Hazardous Zones must be defined for all components of the LNG bunkering supply chain (gas-fuelled ships, bunkering vessel, road tankers and terminal) by their respective owners. This only needs to be performed once as part of the construction process. If changes are made to the equipment, then the Hazardous Zone calculations need to be revised.
The LNG supplier, or the infrastructure owner on the supplier’s behalf, is responsible for checking and confirming the compatibility of the multiple Hazardous Zones between the infrastructure and the gas-fuelled ship. For example, if the zone around an LNG transfer hose impinges on a ventilation intake of the gas-fuelled ship, gas could enter a non-zoned “safe” area. This check must be performed for each new combination of bunkering equipment, location and infrastructure configuration. If each individual zone does not make an acceptable combined zone, bunkering should not proceed.
The checks and calculations should be of sufficient detail to demonstrate to the port/terminal owner that bunkering complies with any regulatory
© Society for Gas as a Marine Fuel 10
consents. The competent authority should review these calculations before granting these consents (permits/regulations) and allowing bunkering. The competent authority should have a right to audit these calculations to ensure compliance.
The port/terminal owner must also ensure that any equipment or personnel that it controls does not permanently or temporarily place ignition sources within the Hazardous Zones that have been defined.
Examples of Hazardous Zones related to bunkering are shown in Table 4.3.
Hazardous Zones can result from small leaks from flanges and valves at any time. Hazardous Zones are therefore relatively small and do not correspond to safety distances resulting from transfers of LNG during bunkering. However, both Hazardous Zones and safety distances require the control of potential ignition sources.
Table 4.3: Hazardous area definitions
Zone Example
Zone 0 / Class 1 Division 1• Only inside the LNG storage tank(s) of a road tanker,
bunker vessel or gas-fuelled ship and in the gas-fuelled ship pipework from storage tank to engine
Zone 1 / Class 1 Division 1 • Inside transfer system and gas-fuelled ship bunkering pipework to the tank
Zone 2 / Class 1 Division 2
• Around any flanged connection on LNG or BOG/vapour return pipework (gas-fuelled ship, LNG bunker vessel and road tanker or terminal)
• Surrounding dry release coupling breakaway flange (also known as an Emergency Release Coupler, or ERC
• Around spindles of all LNG or BOG/vapour-return valves (gas-fuelled ship, LNG bunker vessel and road tanker or terminal)
• Pressure-relief valves that relieve directly to the atmos-phere or the point where vent masts discharge to the atmosphere (gas-fuelled ship, LNG bunker vessel and road tanker or terminal)
© Society for Gas as a Marine Fuel 11
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
4.2. Safety Zone
The Safety Zone can be defined as:The three-dimensional envelope of distances inside which the majority of leak events occur and where in exceptional circumstances there is a recognised potential to harm life or damage equipment/infrastructure as the result of a leak of gas/LNG
The zone is temporary in nature, only present during bunkering. This zone may extend beyond the gas-fuelled ship/LNG road tanker/bunker vessel, interconnecting pipework, and so on, and will be larger than the Hazardous Zone.
The purpose of the Safety Zone is to minimise harm to people and damage to equipment by:
• controlling leaks and spills
• avoiding ignition that could lead to fire or explosion
• excluding non-essential people (to avoid additional injuries or deaths in the event of an accident)
• protecting essential staff through the use of PPE (to minimise the likelihood of injury or death in the event of an accident)
The Safety Zone should always be under control of the PIC and must therefore lie within the port or another entity that allows the PIC the required degree of control.
The size of the Safety Zone will depend on:
• the design of the LNG bunkering infrastructure/gas-fuelled ship
• the configuration of the LNG transfer system
• the amount of LNG inventory involved
• the duration, flow rate and pressure of the potential leak source
• weather conditions and ambient temperature
• the layout of the location where spills could occur
© Society for Gas as a Marine Fuel 12
The owner of the transfer system equipment (normally the LNG supplier/infrastructure owner) should create bunkering scenarios and use these and any relevant mitigations to define – before the first bunkering operation occurs –the safety distances for each combination of bunkering equipment and gas-fuelled ship. Chapter 5 provides guidance on how to do this. The PIC should confirm prior to each bunkering operation that this Safety Zone risk assessment remains correct.
The port/terminal owner should review, agree and accept the safety distances and mitigation strategies provided by the LNG supplier/infrastructure owner (as modified, if necessary, by the gas-fuelled ship) for each bunkering combination of LNG supply infrastructure, flow rate, pressure and so on. These safety distances should be within the operating/bunkering permit provided by the competent authority.
The port/terminal owner, in consultation with the competent (port) authority, will assess the suitability of the proposed bunkering area based on its equipment, operational procedures and the proximity of its staff and the public. If required, the port/terminal owner or competent (port) authority should modify/limit access to the area during the bunkering process as well as reviewing and specifically authorising any SIMOPs in the Safety Zone – be this on the gas-fuelled vessel, within the port/terminal owner’s equipment, or on the LNG supply infrastructure. SIMOPs are covered in SGMF’s report “FP08-01 Simultaneous Operations (SIMOPs) during LNG bunkering”.
Once the area has been defined, the port should provide a means of limiting access to the area to authorised staff through the use of fences/barriers and/or appropriately trained security staff. The procedures required for gaining authorisations for employees, contractors and associated vehicles/equipment should also be communicated.
The port should periodically check the performance of the bunkering operation to check that its requirements and those of the LNG supplier are being complied with. If this is not confirmed, the port should suspend bunkering immediately until compliance is confirmed.
© Society for Gas as a Marine Fuel 13
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Emergency response plans should cover the gas/LNG release scenarios that define the Safety Zone, and any External Zone events required by the competent authority. The LNG supplier/onshore infrastructure owner should prepare a contingency plan to decide which actions to take in each leak scenario and any potential incident escalation (such as fire or explosion). Similarly, the gas-fuelled ship and/or the bunker vessel should prepare a contingency plan to decide their actions in these scenarios. All contingency plans should be developed with the port/terminal owner and port authority, in consultation with the local emergency services.
4.3. Marine Exclusion Zone
The purpose of the Marine Exclusion Zone is to protect the bunkering vessel from other marine traffic, primarily by defining minimum distances and speeds for passing vessels. The Marine Exclusion Zone must also take into account the possibility of collision and determine the potential energy of impact and whether this would damage the bunkering gas-fuelled ship or bunkering vessel in such a way as to cause a major release of gas/LNG.
Definition of the Marine Exclusion Zone is for each port to decide and implement in their port rules based on specific port and ship studies. All ships and bunker vessels should comply with these rules in the normal way.
4.4. Monitoring & Security Area The Monitoring & Security Area is defined as:
The three-dimensional area inside which activities (including people and vehicle movements) need to be identified and monitored to ensure that they do not impact on the safety of the bunkering operation by encroaching on the Safety Zone of the gas-fuelled ship, quayside or LNG bunkering infrastructure
Its primary purpose is to prevent impacts from the actions of others not involved in the bunkering process on the LNG infrastructure or gas-fuelled ship. This could involve unplanned or uncommunicated events such as:
© Society for Gas as a Marine Fuel 14
• vehicles and personnel entering the area
• a ship entering the Marine Exclusion Zone
• the presence of cargo or material that poses an additional risk to the Safety Zone
The area is temporary, only present during bunkering.
Security considerations resulting from the International Ship & Port Security (ISPS) Code are not the only cause of events in the Monitoring & Security Area. In line with the ISPS, the size and restrictions imposed in this zone may be adjusted for short periods at the discretion of the port.
The Monitoring & Security Area will always be larger than the Safety Zone. As the reasons for the Monitoring & Security Area are many and wide-ranging, it is unlikely that it will be possible to define or justify the size of the Monitoring & Security Area by calculation. It should be considered as a contingency on, or factor to, the Safety Zone. This area should only be in place during bunkering operations.
The port should define the nature of the Monitoring & Security Area during bunkering and communicate this, and any options for modifying this area as a result of specific circumstances, to the bunkering infrastructure supplier/bunker vessel (via the LNG supplier) and gas-fuelled ships. This is so that safety, operating and maintenance procedures can be adjusted to allow compliance with the area. Any temporary changes to the Monitoring & Security Area should be communicated as soon as possible to minimise their impact on the bunkering participants.
The Monitoring & Security Area is the responsibility of all the participants in the bunkering process, including the port. This may involve additional manning requirements to perform lookout duties or interventions to divert crossing traffic. Within the Monitoring & Security Area risk assessment of operations will be the norm.
The prime consideration will be about SIMOPs, activities such as cargo loading, passenger movements, maintenance of ship or quayside equipment, and so on, taking place, which, if not correctly controlled,
© Society for Gas as a Marine Fuel 15
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
could encroach into the Safety Zone. SIMOPS are covered in Chapter 7.
Port staff should regularly undertake spot checks, perhaps using checklists, to ensure compliance with their rules by all users of the area and the participants in the LNG bunkering operation.
SGMF has produced separate guidance on SIMOPs.
4.5. External Zone In some jurisdictions – for example, much of Europe – an External Zone may be required. This zone is defined by the amount of risk general members of the public can be exposed to, based on local regulatory requirements.
(Note: The European Union Seveso III 2012/18/EU Directive was developed to safeguard the public from major accident hazards. This is an overarching regulation which is transposed into local law and regulation by each country of the European Union. The Seveso III Directive is for land-based equipment only and does not strictly apply to portable units specified under other legislation, for example ADN and ADR. However, the concept of a societal risk approach remains appropriate.)
A port can insist on training and certain precautions – for example, the wearing of PPE or limited exposure times – for individuals working directly or indirectly in the port, and on visiting ships through its port rules. It cannot influence the way that the public behaves and therefore the level of risk allowed outside the port must be kept low. Risk levels may be modified depending on the number of individuals present – for example, in schools, theatres and places of worship – or by lifestyle limitations in places such as hospitals and prisons, where mobility is limited.
During the planning and authorisation phases of bunkering, the port must consider through safety studies, how far a leak of gas/LNG might reach and the risk this presents to workers and to the public. If local or national criteria for tolerable risk are exceeded, the design or location of the bunkering area/facility must be modified so that the risk criteria are met.
© Society for Gas as a Marine Fuel 16
4.6. Roles of Bunkering Stakeholders in Defining Controlled Zones
The LNG transfer/bunkering process involves at least four stakeholders:
• the LNG supplier and/or bunkering infrastructure owner
• the gas-fuelled ship receiving the fuel
• the owner/operator of the location where the transfer takes place
• the competent authority or authorities (for example, local authority, port authority, flag state, and so on) who define the allowable risks of the bunkering operation and thereby give consent for it to happen
SGMF’s bunkering guide details the roles of all parties at each stage in the bunkering process. Some of the main points appropriate to the control of safety are reiterated here to provide context for the reader. Figure 4.2 shows the relationships between the stakeholders. Tables 4.4 and 4.5 describe the role each stakeholder is expected to play in: firstly, the planning process for bunkering; and, secondly, immediately prior to and during the LNG transfer. Figure 4.2: Stakeholders involved in the regulation and operation of LNG bunkering
Agent (optional)(facilitating the bunkerpurchase and supply)
LNG Supplier(owner/seller
of the LNG)
Bunkering infrastructure owner(owner of the bunkering infrastructure,
eg road tanker or bunker vessel)
Gas fuelled ship (buyer of the LNG)
Class Society(apply
internationalstandards to ships)
Flag State (apply
international regional and
national rules to ships)
Port/Terminal Operator(Operator of a port facility where
the bunkering will take place)
Port Authority(regulator of activities within
a port)
Bunker vessels onlyCompetent Authority
(define risk tolerability which the port authority andbunkering infrastructure owner must comply with)
© Society for Gas as a Marine Fuel 17
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Gas
-Rec
eivi
ng S
hip
LNG
Sup
plie
r/Bu
nker
erPo
rt O
wne
r/O
pera
tor
Com
pete
nt A
utho
rity
HA
ZID
Perfo
rm a
HA
ZID
or s
imila
r stu
dy to
und
erst
and
the
pote
ntia
l haz
ards
and
rang
e of
sc
enar
ios.
Defi
ne ri
sk c
riter
ia re
quire
d.
Haz
ardo
us
Zone
Perfo
rm c
alcu
latio
ns to
de
term
ine
the
haza
rdou
s di
stan
ces
requ
ired
for t
he
ship
-bas
ed L
NG
tran
sfer
sy
stem
s.
Perfo
rm c
alcu
latio
ns to
de
term
ine
the
haza
rdou
s di
stan
ces
requ
ired
for t
he
supp
ly in
frast
ruct
ure
and/
or it
s co
mpo
nent
s.
Perfo
rm c
alcu
latio
ns to
de
term
ine
the
haza
rdou
s di
stan
ces,
if a
ny, r
equi
red
for a
ny p
ort f
acili
ties/
equi
pmen
t.
Revi
ew th
e H
azar
dous
Zon
e ca
lcul
atio
ns, a
nd o
ther
info
r-m
atio
n/da
ta p
rovi
ded.
Mee
t to
agre
e a
com
bine
d vi
ew o
f haz
ardo
us d
ista
nces
and
pre
caut
ions
bas
ed o
n al
l eq
uipm
ent a
nd o
pera
ting
proc
edur
es.
n/a
Safe
ty Z
one
Perfo
rm c
alcu
latio
ns to
det
erm
ine
the
safe
ty d
ista
nces
requ
ired
for t
he L
NG
tran
sfer
.
Mee
t to
agre
e a
com
bine
d vi
ew o
f saf
ety
dist
ance
s an
d pr
ecau
tions
bas
ed o
n al
l eq
uipm
ent a
nd o
pera
ting
proc
edur
es.
Revi
ew th
e Sa
fety
Zon
e ca
l-cu
latio
ns, r
isk
asse
ssm
ents
an
d ot
her i
nfor
mat
ion/
data
pr
ovid
ed.
Mon
itorin
g &
Se
curit
y A
rea
Wor
king
toge
ther
, dev
elop
the
size
and
nat
ure
of th
e M
onito
ring
& S
ecur
ity A
rea
at th
e bu
nker
ing
loca
tion.
Revi
ew th
e M
onito
ring
&
Secu
rity
Are
a pr
opos
al.
Tabl
e 4.
4: S
take
hold
er ro
les
durin
g th
e pl
anni
ng p
hase
© Society for Gas as a Marine Fuel 18
Gas
-Rec
eivi
ng S
hip
LNG
Sup
plie
r/Bu
nker
erPo
rt O
wne
r/O
pera
tor
Com
pete
nt A
utho
rity
Mar
ine
Excl
usio
n Zo
nen/
an/
a
Prep
are
plan
s an
d ru
les
to re
stric
t oth
er s
hipp
ing
pass
ing
dist
ance
s an
d sp
eed
limits
.
Revi
ew th
e M
arin
e Ex
clus
ion
Zone
pro
posa
l.
Exte
rnal
Zo
nen/
aM
eet t
o ag
ree/
unde
rsta
nd p
oten
tial h
azar
d sc
enar
ios
that
mig
ht o
ccur
.n/
a
Emer
genc
y
resp
onse
Dev
elop
an
emer
genc
y re
spon
se p
lan.
Ensu
re th
at a
ppro
pria
te s
kille
d pe
rson
nel,
equi
pmen
t/in
frast
ruct
ure
are
avai
labl
e to
re
spon
d to
the
haza
rd s
cena
rios.
Revi
ew th
e em
erge
ncy
resp
onse
pla
n an
d co
nfirm
th
at s
uffic
ient
cap
abili
ties
are
avai
labl
e.
App
rova
l pr
oces
sn/
a
Prep
are
the
cont
rolle
d zo
ne d
ocum
enta
tion
and
subm
it to
the
com
pete
nt
auth
ority
for a
ppro
val o
n be
half
of a
ll st
akeh
olde
rs.
n/a
Dec
ide
on s
uita
bilit
y fo
r bu
nker
ing
and
docu
men
t an
y ad
ditio
nal l
imita
tions
/re
stric
tions
that
sho
uld
be
impo
sed.
Tabl
e 4.
4 co
ntin
ued:
Sta
keho
lder
role
s du
ring
the
plan
ning
pha
se
© Society for Gas as a Marine Fuel 19
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Tabl
e 4.
5: S
take
hold
er ro
les
durin
g bu
nker
ing
oper
atio
ns p
hase
Gas
-Rec
eivi
ng S
hip
LNG
Sup
plie
r/Bu
nker
erPo
rt O
wne
r/O
pera
tor
Com
pete
nt A
utho
rity
Ope
ratio
ns
Mee
t to
agre
e th
e LN
G tr
ansf
er ra
tes
and
confi
rm th
at
the
bunk
erin
g w
ill b
e w
ithin
the
exis
ting
cons
ents
of t
he
com
pete
nt a
utho
rity.
Port
owne
r/op
erat
or to
be
invi
ted
to p
re-b
unke
r-in
g m
eetin
g bu
t und
er n
o ob
ligat
ion
to a
ttend
.
n/a
Com
mun
icat
e w
ith th
e po
rt, in
goo
d tim
e, to
in
form
it o
f the
tim
e of
the
bunk
erin
g op
erat
ion.
Ale
rt th
e po
rt to
the
tim
esca
le fo
r bun
kerin
g.
App
ly a
ny s
peci
fic v
esse
l tra
ffic
man
agem
ent
proc
edur
es.
n/a
Ope
ratio
nsm
anag
emen
t
Prov
ide
the
pers
on-in
-ch
arge
(PIC
) who
will
op-
erat
e th
e bu
nker
tran
sfer
eq
uipm
ent o
n th
e ga
s-fu
elle
d sh
ip to
ens
ure
the
com
pete
nt a
utho
rity
rule
s/lim
itatio
ns a
re a
dher
ed to
th
roug
hout
the
proc
ess.
Prov
ide
the
pers
on-in
-ch
arge
(PIC
) who
will
op-
erat
e th
e bu
nker
tran
sfer
pr
oces
s an
d eq
uipm
ent
to e
nsur
e th
e co
mpe
tent
au
thor
ity ru
les/
limita
tions
ar
e ad
here
d to
thro
ugh-
out t
he p
roce
ss.
n/a
Perio
dica
lly re
view
the
oper
-at
ions
of t
he L
NG
sup
plie
r/bu
nker
er a
nd/o
r gas
-fuel
led
ship
to e
nsur
e th
at th
ese
com
ply
with
the
term
s of
thei
r bu
nker
ing
cons
ent a
nd, i
f ne
cess
ary,
enf
orce
com
pli-
ance
or r
evok
e th
e co
nsen
t.
Chan
ge
man
agem
ent
Doc
umen
t any
cha
nges
to th
e bu
nker
ing
equi
pmen
t and
/or p
roce
dure
s (in
clud
ing
SIM
OPs
) and
ens
ure
that
any
cha
nges
whi
ch m
ay im
pact
the
bunk
erin
g co
nsen
t are
co
mm
unic
ated
and
app
rove
d by
the
com
pete
nt a
utho
rity.
Perio
dica
lly re
view
the
calc
ulat
ions
/ris
k as
sess
-m
ents
of t
he L
NG
sup
plie
r to
ens
ure
that
thes
e ha
ve
been
am
ende
d to
refle
ct
any
desi
gn, c
ode
or p
ort r
ule
chan
ges.
© Society for Gas as a Marine Fuel 20
SGMF has conducted research to develop a calculation methodology that allows Safety Zone dimensions to be conservatively estimated – reliably and consistently – for a wide range of bunkering configurations, flow rates and locations. The SGMF model, named BASiL (Bunkering Area Safety Information for LNG), is available on SGMF’s website (www.sgmf.info) via the Member Portal.
Existing methods are based on two premises:
• Quantitative Risk Assessment (QRA)
• simple graphs based on defined assumptions – for example, ISO 20519
BASiL is more similar to the QRA approach. It calculates the distance to which flammable gas may extend for a particular failure. QRAs use similar models but provide multiple calculations for a range of variables and then modify these using probabilities. BASiL is deterministic; it does not apply probabilities to the results but uses a representative release size. The correct selection of the appropriate range of variables in BASiL is therefore key – and is the result of the expert judgements of SGMF’s working group.
5.1. Scenario Selection SGMF’s working group undertook a formal HAZID exercise to identify the potential hazardous scenarios clearly. This is reproduced in Appendix B. This working session was deliberately limited to events that might define the safety distance; so it does not include many scenarios that would be considered in a full HAZID. The list was then reviewed to consider what might be appropriate for the definition of the Safety Zone.
A range of failure scenarios was considered, covering both LNG and vapour/BOG leaks. Other scenarios were not considered as they are mitigated by the operation of the Monitoring & Security Area in conjunction with the Safety Zone, require multiple failures, or occur over a such a significant timescale that operator intervention or, in extreme cases, involvement of the emergency services should prevent escalation.
5. Calculating Distances for the Safety Zone
© Society for Gas as a Marine Fuel 21
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
A variety of consequences can occur from leak incidents including
• flash fires
• jet fires
• pool fires
• explosions
• Rapid Phase Transitions (RPTs)
• structural damage resulting from cryogenic spills
Individuals may be harmed – or potentially killed – and property may be damaged by all of these scenarios. However, most have a limited impact area and therefore do not define the size of the Safety Zone. For example, pool fires and jet fires are highly localised so will not define the distance to the edge of the Safety Zone but may define the thermal radiation requirements close to the bunkering ship and infrastructure.
Gas clouds formed by leaking LNG or cold boil-off gas/vapour return can travel significant distances before igniting. The Safety Zone is defined by the maximum distance the gas evaporating from a pool of LNG or from a pressurised LNG release can subsequently be ignited. So the delayed ignition of a gas cloud causing a flash fire is arguably the event that should define the safety distance.
Leaks from the vapour return system (if used), from the vents on the gas-fuelled vessel (vent mast), or from the LNG supply infrastructure are all at relatively low pressure. They therefore have little momentum behind them. As there is limited need for energy from the atmosphere (because they are already gaseous), they disperse in relatively short distances (<10 metres horizontally) compared with LNG leaks, however because they do not stall they can reach greater heights.
Gas clouds produced from – vaporising – LNG leaks therefore define the extent of the Safety Zone.
© Society for Gas as a Marine Fuel 22
5.2. Failure Scenarios After extensive consultation with members and cryogenic industry bodies, SGMF has yet to identify a failure mechanism where a hose or loading arm rupture, where the failure creates a hole so large that it covers most of the surface area of the hose or breaks the hose in two (also called a guillotine failure), is likely to occur – with one exception. All evidence from industry suggests that hoses fail before rupturing, allowing the transfer to be stopped and the hose removed from service and disposed of.
The only likelihood of a hose or loading arm rupturing is from a vessel or road tanker pulling away with the transfer system still attached. The IGF Code has a requirement for a breakaway device that will split the loading arm/hose in a controlled manner and prevent/minimise LNG outflow. A guillotine failure therefore is a double jeopardy event: firstly tow away and secondly failure of the breakaway valve. Two failures with dissimilar causes are not normally considered in safety assessments. In the light of this, a hose rupture is not considered here.
Some experimental work has been performed looking at hose damage and failure. Too few experiments have been performed for the results to be conclusive but they have shown that cryogenic hoses can suffer significant levels of damage – for example, being run over by a LNG road tanker – without failing. In all instances the damage to the hose was obvious and any hose inspection prior to use, as required by the IGF Code and SGMF’s bunkering guidelines, should result in the hose being rejected before use.
SGMF has therefore placed emphasis on determining a representative hole size for a transfer system failure. Within a Quantitative Risk Assessment (QRA) a range of release sizes would be analysed and appropriate frequencies assigned to each case. For the purposes of defining the Safety Zones within this study, a deterministic approach is taken which requires the selection of a single representative release size for a given transfer operation. This approach is consistent with the “Deterministic assessment of the Safety Zone” described in ISO 20519.
The LNG industry can use only two types of hose:
© Society for Gas as a Marine Fuel 23
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
• a metallic hose consisting of a flexible metal tube supported by an internal and external wire to give the hose the necessary pressure performance
• a composite hose which consists of layers of LNG impermeable fabrics and insulation, again wrapped in a structural wire for mechanical strength
Most failure data are for rubber hoses used for oil and LPG. These are not appropriate for LNG.
Metallic hoses tend to be used for smaller sizes as they are lighter and more easily handled. Composite types are generally used for larger transfer hoses.
SGMF has collated a variety of experience (on a confidential basis) from the filling of LNG road tankers and cryogenic gas (liquid nitrogen, liquid oxygen) operations to establish good practice. About five million LNG truck transfers are estimated to have occurred and the cryogenic gases industry should exceed this by at least an order of magnitude. The findings of this work are:
• no guillotine failures have been recorded
• 3,000-4,000 transfers through a hose have been achieved without leakage during the filling of LNG road tankers
• hose handling determines hose life; the main cause of failure is fatigue from temperature, pressure and movement cycles, particularly bending
• fatigue failures produce small leaks, at least initially, which are readily detectable, allowing time for the hose to be removed from service and destroyed
• small leaks can occur through poor hose management, such as inadequate leak checking, worn couplings, damaged gaskets, dirt and debris on flange faces and gaskets and the errors and bad habits of operators
Loading arms can be treated as fixed pipework and fittings and so are much easier to quantify using industry databases.
Additional details of SGMF’s assessment can be found in Appendix C.
© Society for Gas as a Marine Fuel 24
5.3. Leak Behaviour Many factors determine how far a gas cloud will spread and remain within flammable concentration limits. The only way of fully assessing all the possibilities is to conduct a Quantitative Risk Assessment (QRA). This section describes a simplified technique that provides a conservative estimate of the required safety distances. In simplifying the calculation process some of the less important phenomena are not assessed but are covered by reasonable assumptions.
As anticipated and evidenced by the different safety distances for refuelling LNG-fuelled trucks, compared with bulk LNG transport and unloading, size matters. The greater the transfer rate, the larger the transfer equipment, the larger the potential leak and therefore the larger the safety distance. So a single safety distance will not apply to all ships; if it works for a large container ship it would be hugely conservative for a small Ro-Ro ferry, and, if based on an Offshore Support Vessel (OSV), would be far too small to work for a cruise liner.
The parameters considered by SGMF include:
• different orientations of leaks – vertically upwards, horizontally and downwards
• LNG transfer flow rate, temperature and pressure
• hole size
• a range of LNG compositions and physical properties
• various climatic conditions around wind speed, climatic stability, ambient temperature and humidity
• various durations of release (for different types of ESD systems)
• different geometries/topographies for releases over land and water and at different elevations
Importantly, these parameters interact with each other. This means that some effects must be considered together, increasing the model complexity and resulting in the need to consider about 1.4 million data points for eight parameters. Each parameter is described below:
© Society for Gas as a Marine Fuel 25
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
5.3.1. Leak OrientationThree leak orientations define LNG behaviour: vertically upwards, horizontal and vertically downwards. Leaks at all angles in between can be shown to travel less far than these orientations.
Vertically upward releases initially flow upwards before eventually stalling, returning to the water/land surface and spreading out. The vertical extent of the Safety Zone defined by these releases is sensitive to density and the initial velocity of the release.
Horizontal releases maintain their horizontal momentum, with density effects causing the plume to drop towards the ground or water surface.
Downward releases impact on the ground or the water surface, losing most of their momentum. Liquid in the jet forms a pool on the surface, and the vapour evaporating from this pool forms a low-momentum dispersing plume.
5.3.2. LNG Transfer Flow Rate, Temperature and PressureThe larger the transfer rate, the larger the transfer equipment, the larger the potential leak and therefore the larger the safety distance.
Transfer pressure and temperature define, along with the hole size, the initial momentum and so play an important role for horizontal and vertical distance calculations, significantly modifying those based on flow rate alone.
The temperature and pressure of the LNG in storage prior to transfer (that is, pumping to transfer pressure) define the energy in the LNG and how much of the LNG vaporises (“flashes”) at the leak point and how much remains as a liquid. To vaporise the liquid LNG must absorb energy from its surroundings.
5.3.3. Hole SizeReliability assessment work on piping in the offshore oil and gas industry – for example, the Process Leak for Offshore installation Frequency Assessment Model (PLOFAM) project – and for failure in rubber LPG/oil
© Society for Gas as a Marine Fuel 26
hoses – for example, the Dutch RIVM methodology – links hole size to pipe/hose size. SGMF has adopted the same approach for its guidance on metal and composite hoses.
The only failure size advice available for metal hoses is from the UK Health & Safety Executive and is for metal hoses transferring chlorine. Its expert judgement was that hole diameters of 3 mm and 15 mm were appropriate for calculating safety distances. Applying these holes sizes across the transfer hose sizes used in the LNG bunkering industry suggests that a hole size of 6% of the hose diameter is appropriate. Industry experience suggests this is conservative.
There is less information available about the failure of composite hoses. Manufacturers suggest failures in composite hoses produce one or more 1 mm holes or a short tear, a few millimetres long. If the perimeter of a tear is considered to be the circumference of a hole, similar hole sizes to metal hoses can be suggested. The conservatism of the model is therefore maintained.
If the design of the hose is such that the likelihood of releases from the hose itself is low, then releases from the flanges and valves at either end of the hose will dominate the release distance. In this case the Safety Zone would be based on releases from these locations.
SGMF suggests that the following hole sizes are appropriate for estimating the size of the LNG/gas cloud/pool and therefore the size of the Safety Zone.
© Society for Gas as a Marine Fuel 27
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Table 5.1: Hole sizes used in BASiL
Size Metal or composite hose
Fitting/gasket/valve, fixed piping or hard arm
2 inch / 50 mm 3 mm 3 mm
3 inch / 75 mm 4.5 mm 4 mm
4 inch / 100 mm 6 mm 4 mm
6 inch / 150 mm 9 mm 4 mm
8 inch / 200 mm 12 mm 5 mm
10 inch / 250 mm 15 mm 5 mm
Basis Hole diameter is 6% of hose diameter
PLOFAM method using a 15 m hose, 8 valves and 11
flanges
The hole diameters shown are circular, however, a hydraulically equivalent diameter can be calculated from any hole geometry by equating the actual hole perimeter length to the circumference of a circle.
These sizes are acknowledged to be conservative. SGMF continues to look at ways to improve hole size estimates through dialogue with the industry and the sourcing of data and experience.
These hole sizes are based on the following assumptions:
• the Safety Zone is under the control of the Person In Charge (PIC) of the bunkering; should the PIC or another operative identify a leak, the system will be shut down immediately
• management systems have been developed and implemented to ensure that damaged hoses are identified prior to use through inspection
• if a leak is found during inspection or during the bunkering operation and cannot immediately be sealed, the hose will be immediately removed and destroyed
• the scenario of a ship and bunker facility (vessel or road tanker) moving away from each other is controlled both by procedures and
© Society for Gas as a Marine Fuel 28
the requirement of the IGF Code for a dry-break coupling, which is a self-sealing weak point in the transfer system designed to eliminate/minimise spillage
• the potential for other impact damage is controlled by appropriate procedures, particularly in relation to allowable operations being conducted simultaneously with LNG bunkering (SIMOPs)
5.3.4. LNG Compositions and Physical PropertiesLNG composition has a minor impact on safety distances, primarily through its effect on the calculation of flammability limits (LFL and UFL) and on the level of initial flashing as the LNG exits a hole.
5.3.5. Atmospheric Conditions Atmospheric conditions can significantly impact gas dispersion distances. So, to be effective, any analysis must include climatic factors based on the bunkering location. Temperature and humidity are primarily a function of latitude but analysis has shown significant scatter unless a longitude component is also included.
Stable atmospheric conditions and low wind speeds, often found at night (a Pasquill stability factor of F), lead to reduced turbulence and therefore the largest safety distances. These conditions occur only rarely. Use of the F factor is justified for the External Zone but not for the Safety Zone, where a D Pasquill stability factor – the most conservative daytime option – has been assumed.
5.3.6. Various Durations of ReleaseThe duration of a release depends on how long it takes to detect and isolate the leak. The IGF Code requires an Emergency Shut-Down (ESD) system to be used during bunkering. To be effective, the ESD system must act faster than an LNG/gas leak achieves its maximum extent. SGMF has included four ESD scenarios in the model:
• a fully automatic, fast acting ESD system which acts in 10 seconds including leak detection and valve closure
• a fully automated ESD system which acts in 30 seconds (from leak detection to full valve closure)
© Society for Gas as a Marine Fuel 29
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
• a semi-automatic ESD system which requires some level of human interaction – for example, responding to an alarm and hitting the ESD button or manually closing a valve – and therefore takes longer to implement; two minutes has been selected as this is a common value in oil, gas and chemicals systems where similar activities take place
• a fully manual ESD system which requires multiple human actions to occur, where a 10 minute period is commonly employed
5.3.7. Different Geometries/Topographies for Releases Over Land and Sea and at Different ElevationsIf a hose is on the ground or on some other solid surface, air cannot get beneath the dispersing gas cloud so air entrainment is considerably reduced, leading to larger safety distances. The effect is very specific; even minor changes in elevation up to perhaps 1 m make considerable differences. Above 1 m the effect is negligible. The LNG or cold vapour behaves differently depending on whether it impacts water, soil/concrete or the steel hull of a vessel.
5.4. The SGMF Safety Distance Model – BASiL SGMF has created a model called BASiL to estimate the size of the Safety Zone based on the extent of the gas cloud to 100% LFL.
All the parameters described previously need to be considered simultaneously as they all interact and in different combinations lead to different safety distances.
BASiL consists of a database of 1.4 million combinations of the eight input parameters:
1. location (latitude and longitude)2. amount of LNG transferred and the duration of the transfer3. LNG supply temperature and pressure4. net Calorific Value (composition)5. transfer pressure6. transfer system elevation (above solid surface)7. hose/transfer system diameter8. ESD method
© Society for Gas as a Marine Fuel 30
The model interpolates simultaneously within the most appropriate points in the database to derive estimates of these sizing parameters for the Safety Zone.
The following parameters are not considered in BASiL:
• tank size, type and location, because failures here could happen at any time and are not specific to bunkering
• SIMOPS, because safety distances influence which SIMOPs are allowed and their location not vice versa (see Section 7 for details)
• the geography of the port, because the parameters are too wide to cover; the methodology looks at an “open” area as this gives the most conservative assumption; features of ports which block this free movement of gas should be conservative but, if appreciable, would need to be covered by more detailed risk assessment and consequence calculations
• the frequency of ship movements, covered by the Marine Exclusion Zone
Users interact with BASiL through a web page. The web form allows the user to enter the calculation parameters and review the results without being able to compromise the underlying dispersion distance database.
The data required to calculate the safety distances and the limits of the parameter ranges modelled are described in Table 5.2
Table 5.2: BASiL data requirements
Data Explanation
Bunkering information and date This enables the user to record the case being calculated.
Volume or mass of LNG transferred and time scale of transfer
Determines the transfer flow rate. Used directly by the software to determine leak behaviour
Latitude and longitude
The nearest port maximum temperature and relative humidity found from a port list using latitude and longitude. Latitudes north of the Equator are positive, south are negative. Longitudes east of Greenwich (London) are positive, west are negative.
Used directly by the software to determine safety distances.
© Society for Gas as a Marine Fuel 31
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Data Explanation
Supply typeRoad tanker, shore terminal or bunker vessel.
Only used by the software to change the output image.
Supply storage pressure and temperature
1 to 12 barg and -170 to -115°C.
If the supply storage temperature is not known then the LNG is assumed to be saturated and the saturation temper-ature is calculated from the supply storage pressure.
If the supply storage temperature is known then this is used as the LNG temperature
These values are used to determine safety distances via in-termediate variables (interpolation used on lookup tables).
LNG Net Calorific Value (LHV) and LNG density
34 to 40 MJ/Nm3 (at 15°C/15°C, 1.01325 bar).
Used as a proxy for LNG composition.
Used directly to determine safety distances (interpolation used on lookup tables). If not known, LNG density is calculated from the calorific value
Transfer pressureGreater than supply storage pressure and less than 20 barg.
Used to determine safety distances via intermediate varia-bles (interpolation used on lookup tables).
ESD activation
Fully automatic (control system linked), semi-automatic or fully manual.
Used to select one of four sets of results for dispersion from pools corresponding to ESD times of 10 seconds, 30 seconds, 2 minutes and 10 minutes.
Primary leak source
Hose or fitting.
Release diameters are shown in Table 5.1.
Release diameter is used to determine safety distances via intermediate variables (interpolation used on lookup tables).
Maximum hose elevation above a continuous solid surface (eg ground or deck)
0-5 m.
If greater than 5 m then use 5 m.
Used to determine safety distances (interpolation used on lookup tables).
Transfer system/hose internal diameter
2-10 inch diameter.
Used to determine release diameter if primary leak source is a hose.
© Society for Gas as a Marine Fuel 32
Table 5.2 continued: BASiL data requirements
Data Explanation
Bunkering Station Layout
Open deck or semi-enclosed bunker station.
Safety zone image changes to indicate hose entry location.
If semi-enclosed is selected there are no pool releases on the deck and the extent of the jet zone on the deck is a function of the Distance Below Deck (see next row).
Distance Manifold Below Deck
0-20 m.
Used to determine the extent of the jet zone on the deck.
If Distance Below Deck > H1 then the jet zone does not extend onto the deck.
This result appears as an additional result below the stand-ard Safety Zone distances.
An example of the BASiL data input and output screen is shown in Figure 5.1.
Figure 5.1: Example BASiL calculation for truck-to-ship bunkering
BASiL produces up to three sets of safety distances, one for each orientation of leak, along with a graphic showing what these mean for the bunkering method selected:
Note: BASiL image is indicative and may change as the model develops
© Society for Gas as a Marine Fuel 33
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
• horizontal and vertically upwards dispersion (H1 and R1)
• downwards dispersion onto land (H2 and R2) and onto water (H3 and R3)
• if the bunkering location is below deck, an additional distance – R4 – is shown which indicates how much of a vertically upwards release reaches the main deck and disperses across it
A close-up of the bunkering results for a road tanker and bunkering vessel is shown in Figures 5.2 and 5.3. Table 5.3 shows the full definition of the safety distances.
Figure 5.2: Bunkering from a bunker vessel to a manifold located in the ship’s side
Note: BASiL image is indicative and may change as the model develops
© Society for Gas as a Marine Fuel 34
Figure 5.3: Bunkering from a road tanker to a manifold on the ship’s deck
Note: BASiL image is indicative and may change as the model develops
© Society for Gas as a Marine Fuel 35
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Rele
ase
Dis
tanc
eA
pplic
atio
nPl
an/e
leva
tion
3D v
iew
Hor
izon
tal
jet
Radi
us
(hor
izon
tal)
from
hos
eR1
Radi
us is
app
lied
hor
izon
-ta
lly a
long
leng
th o
f tra
nsfe
r sy
stem
and
at e
nds
of h
ose.
Verti
cally
up
war
ds
jet
Hei
ght a
bove
ho
seH
1
Hei
ght i
s ap
plie
d ab
ove
the
heig
ht o
f hos
e ov
er
the
horiz
onta
l ext
ent o
f the
Sa
fety
Zon
e de
fined
by
the
horiz
onta
l rad
ius*
.
Dow
nwar
d re
leas
e on
to la
nd/
deck
Radi
us
(hor
izon
tal)
from
hos
eR2
Radi
us is
app
lied
horiz
on-
tally
alo
ng le
ngth
of h
ose
abov
e th
e la
nd o
r shi
p de
ck
and
at e
nd o
f hos
e.
Hei
ght a
bove
la
nd o
r shi
p de
ckH
2
Hei
ght i
s ap
plie
d ab
ove
the
land
or s
hip
deck
ove
r th
e ho
rizon
tal e
xten
t of t
he
Safe
ty Z
one
defin
ed b
y th
e ho
rizon
tal r
adiu
s.
Dow
nwar
d re
leas
e on
to w
ater
Radi
us
(hor
izon
tal)
R3
Radi
us is
app
lied
horiz
on-
tally
at e
nd o
f gap
bet
wee
n sh
ip a
nd la
nd o
r shi
p an
d bu
nker
ves
sel.
Hei
ght a
bove
w
ater
H3
Hei
ght i
s ap
plie
d ab
ove
sea
over
the
horiz
onta
l ext
ent o
f th
e Sa
fety
Zon
e de
fined
by
the
horiz
onta
l rad
ius.
Tabl
e 5.
3: In
terp
reta
tion
of B
ASiL
resu
lts
* N
ote
that
Saf
ety
Zone
ext
ends
bel
ow th
is to
eith
er th
e la
nd, s
hip
deck
or w
ater
© Society for Gas as a Marine Fuel 36
The database within BASiL is based on calculations from specialist gas dispersion models which were derived from large-scale experiments conducted using LNG by British Gas (now DNV GL) as part of joint industry projects. The experimental dataset is the basis of most LNG consequence models used in QRAs.
Reviews of accuracy against the dispersion models have shown that the vast majority of the BASiL calculation result in safety distances within ±10% of the actual values (Appendix E). LNG composition appears to be the source of most of the errors where a single parameter, Net Calorific Value (also known as Lower Heating Value), is being used to cover several different effects.
5.5. Example Calculations
BASiL predicts safety distances ranging from 7 m to 140 m in the horizontal direction, primarily based on transfer flow rate/pressure considerations, and 3 m to 45 m in the vertical direction.
Example 1: Bunkering a Container ShipThis example covers the fuelling of a 10,000 TEU (Twenty Foot Equivalent Unit) container ship with 2000 m3 of LNG over a 5 hour period from an 8 inch hose using LNG supplied by pump from a bunker vessel at 6 bar. The container ship’s manifold is located in the hull (semi-enclosed), beneath the accommodation, 10 m below the main deck. The hose has a minimum height above the deck of the bunker vessel and the sea surface of 5 m. The ESD system is semi-automatic, acting within 2 minutes.
The safety distances for four container ports are shown in Table 5.3.
© Society for Gas as a Marine Fuel 37
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Table 5.3: Container ship example
Port
Horizontal jet Vertical jet Downwards over water Across
deck
Radius (R1), m
Height (H1), m
Radius (R3), m
Height (H3), m
Radius (IR1), m
Los Angeles, USA 36 16 48 5 13
Rotterdam, the Netherlands 36 15 47 5 12
Shanghai, China 36 16 49 6 13
Singapore 37 16 52 6 14
Taking Singapore as the example location, the Safety Zone is shown in Figure 5.4 by the pink area which reaches around the hull of the container ship. The height that the cloud reaches before stalling means that some of the gas may flow across the main deck (in red). Depending on the location of the bunkering manifold, up to two stacks of containers (of 175 in this example) can lie within the Safety Zone. Safety zone procedures should limit activities in this area, such as removing and loading containers, and operations of reefer containers (non-intrinsically safe electric motors and storage of hazardous cargo containers) unless a risk assessment says otherwise. The rest of the container ship, the other 18 rows, could be loaded and unloaded as normal unless these loading operations pass through the Safety Zone.
Figure 5.4: Safety zone contours for Singapore example
Safety
Bunkering safety zoneOn ship’s main deck 10,000 TEU Container Ship
LNG bunker vessel
LNG bunker manifold
zone
© Society for Gas as a Marine Fuel 38
There is little variation between the distances based on port location (temperature, humidity, worst case wind/atmospheric stability).
Example 2: Bunkering a Ro-Ro FerryThis example covers the fuelling of a Ro-Ro ferry with 200 m3 of LNG over 2 hours from a 4 inch hose using LNG supplied by pump from a road tanker filled at 2.5 barg and pumped to 6 bar. The ferry’s manifold is located in the hull, 10 m below the main deck. The hose has a minimum height above the ground of 0.1 m. The ESD system is semi-automatic, acting within 2 minutes.
The safety distances for seven ferry ports are shown in Table 5.4 and an example for New York in Figure 5.5.
Table 5.4: Ro-Ro ferry example
Port
Horizon-tal jet
Vertical jet
Downwards over water
Downwards over land
Across deck
Radius (R1), m
Height (H1), m
Radius (R3), m
Height (H3), m
Radius (R2), m
Height (H2), m
Radius (IR1), m
Buenos Aires, Argentina 27 9 25 3 20 3 0
Chittagong, Bangladesh 27 10 26 4 21 4 0
Hong Kong 27 10 27 3 21 4 0
Manilla, the Philippines 27 10 27 3 21 4 0
New York, USA 27 9 25 3 20 3 0
Piraeus, Greece 27 9 25 3 20 3 0
Stockholm, Sweden 27 9 25 3 20 3 0
© Society for Gas as a Marine Fuel 39
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Figure 5.5: Ro-Ro ferry example
Manifold
2 x LNG road tankers
Vehicle loading at stern Ro-Pax ferry
Safety Zone
120 m
Again, there is little variation between the distances based on port location (temperature, humidity, worst case wind/atmospheric stability).
The Safety Zone does not reach the main deck of the ferry and, as located, loading of vehicles is significantly outside the Safety Zone.
A variety of step-off cases were modelled, using New York as the base case, to see what reductions in safety distances would be possible by varying the main parameters, as shown in Figure 5.6.
© Society for Gas as a Marine Fuel 40
Figure 5.6: Tornado plots of main parameter effects
0 5 10 15 20 25 30 35
Hose elevation
ESD system
Transfer pressure
Hose size
Combination
A: Horizontal jet
0 2 4 6 8 10 12
Hose elevation
ESD system
Transfer pressure
Hose size
Combination
B: Vertical upwards jet
0 10 20 30 40
Hose elevation
ESD system
Transfer pressure
Hose size
Combination
C: Vertically downwards jet
0 5 10 15 20 25 30 35
Hose elevation
ESD system
Transfer pressure
Hose size
Combination
A: Horizontal jet
0 2 4 6 8 10 12
Hose elevation
ESD system
Transfer pressure
Hose size
Combination
B: Vertical upwards jet
0 10 20 30 40
Hose elevation
ESD system
Transfer pressure
Hose size
Combination
C: Vertically downwards jet
0 5 10 15 20 25 30 35
Hose elevation
ESD system
Transfer pressure
Hose size
Combination
A: Horizontal jet
0 2 4 6 8 10 12
Hose elevation
ESD system
Transfer pressure
Hose size
Combination
B: Vertical upwards jet
0 10 20 30 40
Hose elevation
ESD system
Transfer pressure
Hose size
Combination
C: Vertically downwards jet
Size of Safety Zone in m
Size of Safety Zone in m
Size of Safety Zone in m
Note: The green bar represents the possible reduction in the Safety Zone by moving to the next lower value of a parameter. The red bar represents the increase in the Safety Zone from the next higher value.
The base case is the boundary between green and red.
© Society for Gas as a Marine Fuel 41
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
The horizontal and vertically downwards jets determine the safety distances. The biggest effects on the size of the safety distance are the impact of the transfer pressure and hose size. The best combination of parameters leads to a reduction in the size of the Safety Zone by about half.
Table 5.5: Sensitivity study results
Port
Horizon-tal jet
Vertical jet
Downwards over water
Downwards over land
Across deck
Radius (R1), m
Height (H1), m
Radius (R3), m
Height (H3), m
Radius (R2), m
Height (H2), m
Radius (IR1), m
Base case 27 9 25 3 23 3 0
Best combination 9 5 11 2 13 2 0
© Society for Gas as a Marine Fuel 42
6. Calculating Distances for the Monitoring & Security Area The Monitoring & Security Area is not controlled but monitored. The aim is to stop issues occurring in the area from affecting the Safety Zone.
These could include:
• unauthorised vehicles or vessels
• unauthorised people
• ignition sources
No activity is prohibited in the Monitoring & Security Area so long as precautions have been identified and taken to limit any possible impact on the Safety Zone and required emergency responses are clear. For example, ignition cannot be excluded and protocols to control access from the Monitoring & Security Area to the Safety Zone must be present and enforced, either by staff or fencing and signage, or both. Personnel working in the Monitoring & Security Area should have been briefed (or trained) to understand the special features of this area. Therefore the port/terminal needs to assume some control and monitoring of the area. Formal risk assessment of activities in the Monitoring & Security Area could be performed but in many instances it may be better for the PIC and/or port/terminal supervisor to just continuously risk assess during the bunkering operation as they would normally do.
The extent of the Monitoring & Security Area is therefore limited to port-controlled areas which can be observed, either directly or via CCTV or similar systems (if visible to the PIC). The reaction time to potential threats defines the minimum extent of the area. Some risks may be port-wide – for example, extreme weather or civil disobedience – whose onset is slow and can be managed with a relatively small Monitoring & Security Area. Other events will be very localised and require either quick reaction times or extensive Monitoring & Security Areas. The Marine Exclusion Zone performs the same function as the Monitoring & Security Area if bunkering takes place away from the quayside. The Monitoring & Security Area can extend into the External Zone provided monitoring is possible.
© Society for Gas as a Marine Fuel 43
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
The PIC should watch the Monitoring & Security Area and interpret any movements or actions that might interfere with the Safety Zone. If a threat is confirmed or looks likely to be confirmed, and attempts at communicating with individuals involved have proven unsuccessful, the PIC should use their awareness of the hazards and emergency plans to decide whether to suspend or shut down the LNG transfer, make the area safe and raise the alarm with the port/terminal operator.
Similarly, as the Monitoring & Security Area is linked with the International Ship and Port Facility Security (ISPS) Code, the PIC may decline to start, suspend or shut down the LNG transfer if he/she feels that control of the bunkering process has been or could be lost.
Relocation of the bunkering process is possible but unless the back-up area has also been risk-assessed it probably only makes sense to relocate if external events are significant and have a long onset period.
In ports where large numbers of people are present, be they passengers or workers, it is the responsibility of the port/terminal to look after them. As the External Zone only covers areas outside the control of the port/terminal, their welfare should be considered under the Monitoring & Security Area guidelines.
© Society for Gas as a Marine Fuel 44
7. SIMOPS and Non-Standard OperationsSIMultaneous OPerations (SIMOPs) can add additional hazards, particularly ignition sources, and also may distract those involved in bunkering. From a port perspective, therefore, SIMOPs should only be allowed if it can be demonstrated that they do not create an unacceptable level of risk.
The Safety Zone is defined as an area where special precautions are taken to limit access, ignition sources and other hazardous events. So SIMOPs must not be allowed to compromise the Safety Zone. Extensive risk assessment is required to justify SIMOPs within the Safety Zone.
SGMF understands that SIMOPs are normally a commercial necessity. However, they must not significantly impact the safety of the LNG transfer process. SGMF believes that procedures and rules can be designed to successfully allow SIMOPs, through co-ordination between the competent authority, the terminal operator, the bunker infrastructure owner and the gas-fuelled ship.
There are four general types of SIMOPs, as shown in Table 7.1.
Table 7.1: SIMOPs categories
Regular SIMOPs Non-standard but planned
Operations that happen in the same or very similar way on a frequent basis, such as people/passenger/crew movements and cargo loading and unloading.
Operations that happen infrequently but are known and can be planned for. For example, maintenance or lifeboat drills.
Non-standard and unplanned External activities
Operations that occur unexpectedly and infrequently and need immediate atten-tion. For example, breakdowns.
Activities or events, normally short-term and irregular, that are beyond the control of the bunkering stakeholders and potentially the terminal/port. For example, security alerts or public festivals.
© Society for Gas as a Marine Fuel 45
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Regular SIMOPs can be risk-assessed and approved well in advance of the bunkering taking place, via the bunker plan/management system.
It is important to consider regular SIMOPs at the design stage. Location of the bunkering manifold can then be optimised to be away from cargo operations and passenger movements or to include ventilation systems to prevent flammable gases entering the ship or the build-up of flammable atmospheres. Another example would be bunkering container ships while loading cargo. The safety of the bunker vessel should not be compromised by a dropped container (or stack of containers) which may require changes to loading priorities/sequences. Most non-standard operations can be envisaged to occur at some stage so can be risk assessed in advance. Risk assessment immediately prior to bunkering is strongly discouraged. If it must take place, it must involve multiple individuals from all bunkering stakeholders who should unanimously agree that the SIMOP does not create an unacceptable level of risk.
Tankers carrying flammable and/or toxic oil products and chemicals have procedures and cultures for safe working in their cargo zones. These vessels should easily be able to provide the necessary documentation to justify SIMOPs. Other ship types – for example, container ships – do not have this culture and therefore may find it more onerous, at least initially, to demonstrate that they can safely perform SIMOPs.
SIMOPs are location, bunkering system and vessel specific.
If SIMOPs cannot be agreed, bunkering should not take place at the same time as these other activities.
The PIC must stop bunkering if any SIMOPs impact, or appear likely to impact, any part of the bunkering process or its equipment. The gas-fuelled vessel’s Captain/POAC has the same responsibility from the ship side. Port enforcement officers may be present and should also be able to stop bunkering.
SGMF has published detailed guidance on SIMOPs in publication FP08-01 “Simultaneous operations (SIMOPs) during bunkering”.
© Society for Gas as a Marine Fuel 46
8. Bibliography
8.1. Publications
Society for Gas as a Marine Fuel, “gas as a marine fuel, an introductory guide”
Society for Gas as a Marine Fuel, FP07-01 “safety guidelines, bunkering”
Society for Gas as a Marine Fuel, FP04-02 “Bunkering of ships with Liquefied Natural Gas (LNG), competency and assessment guidelines”
Society for Gas as a Marine Fuel, FP08-01 “Simultaneous Operations (SIMOPs) during LNG bunkering”
Lloyds Register, “Process leak for offshore installations frequency assessment model – PLOFAM”, report number 105586/R1, 2016
8.2. Relevant International Standards
8.2.1. International Maritime OrganizationInternational Code for the Construction and Equipment of Ships Carrying Liquefied Gases in Bulk (IGC Code)
International Code of Safety for Ships using Gases or other Low flashpoint Fuels (IGF Code)
International Safety Management Code (ISM Code)
International Ship & Port Security Code (ISPS Code)
International Dangerous Goods (IMDG) Code
8.2.2. European Union Directives and StandardsSeveso III (2012/18/EU) Directive
ATEX (99/92/EC & 94/92/EC) Directives
ADN – Accord Européen Relatif au Transport International des Marchandises Dangereuses par Voies de Navigation Intérieures (European Agreement Concerning the International Carriage of Dangerous Goods by Inland Waterways)
ADR – Accord Européen Relatif au Transport International des Marchandises Dangereuses par Route (European Agreement Concerning the International Carriage of Dangerous Goods by Road)
© Society for Gas as a Marine Fuel 47
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
8.2.3. ISO and Other StandardsISO 17776 Petroleum and natural gas industries – Offshore production
installations – Guidelines on tools and techniques for hazard identification and risk assessment
ISO 18683 Guidelines for systems and installations for supply of LNG as fuel to ships
ISO 20519 Specification for bunkering of gas-fuelled ships
IEC 60079 Explosive Atmosphere Standards
IEC 60092 Electrical installations in ships – Tankers – Special features
NFPA 70 National Electrical Code
© Society for Gas as a Marine Fuel 48
Appendix A: Hazards Associated With LNGA1. LNG LNG is a mixture of hydrocarbons – predominately methane (80-99%) – in a liquid state. Other significant components are other alkanes: ethane, propane and butane. Nitrogen may also be present at a concentration of up to 1%.
LNG is a liquid fuel produced from natural gas. At normal conditions, atmospheric pressure and ambient temperature natural gas exists only in the gaseous phase. Liquefied gases are produced by reducing their temperature and/or increasing their pressure until the gas liquefies. Small amounts of liquid can evaporate into very large volumes of gas. For example, one litre of LNG vaporises to about 600 litres of natural gas when warmed to ambient temperatures.
Bulk LNG is normally stored at atmospheric pressure. To be a liquid at this pressure, natural gas must be cooled to about -162°C (or -262°F). On LNG-fuelled ships LNG is stored at higher pressures, typically up to 4-5 barg. At these pressures LNG remains liquid at higher temperatures. However, all of these temperatures are very cold – for example, the boiling point of LNG at 4 barg (5 bara) is -138°C.
As a liquid, LNG is relatively non-hazardous, neither flammable nor toxic. It is also colourless and odourless. LNG’s cryogenic temperature does present hazards in that personnel can be harmed and most materials, including ship hull steels, can be damaged if they accidentally come into contact with LNG. This, however, is a localised effect and will not define safety distances.
No matter how well an LNG’s container is insulated, some heat will transfer into the LNG, causing some of it to vaporise. This vapour is flammable so its behaviour defines the required size of the Hazardous Zone and the Safety Zone.
© Society for Gas as a Marine Fuel 49
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
A2. Boil-Off Gas and Natural Gas Vapour generated from evaporating LNG is called Boil-Off Gas (BOG). This vapour is very similar to natural gas (NG). The phrase “vapour return” denotes the gas in a LNG storage tank that is being displaced by LNG during bunkering and flowing from the gas-fuelled ship back to the LNG supplier. Vapour return is just BOG flowing from one tank to another.
BOG, being natural gas, is colourless and odourless. Pipeline natural gas has a smell added to allow leaks to be easily detected. This is not so with LNG. The odorant used for pipeline natural gas would freeze if put into the LNG and be ineffective.
Because BOG and natural gas are flammable, their behaviour can define the required size of the Hazardous Zone and the Safety Zone.
A3. LNG Leak Sources
LNG or BOG can escape into the atmosphere in several ways. The most obvious are through a leak during LNG transfer or poor management of the BOG generated during LNG transfer. Bunkering is particularly susceptible to leakage because temporary connections have to be made between pipes or hoses. These connections must be continually checked for leak tightness throughout the bunkering process. The coldness of the LNG causes the connectors and hoses/pipes to contract, which may cause connectors to loosen and then leak. The gas-fuelled vessel may also move relative to the bunkering infrastructure which can put strain on the connectors or even on the hose/pipework itself, causing leaks.
Several scenarios arise during bunkering which could potentially generate significant amounts of BOG. If these are not handled correctly by the PIC, BOG may be released to the atmosphere to prevent larger LNG spills resulting from over-pressurisation or overfilling of equipment. Venting gas to the atmosphere is not allowed except in emergency conditions. Vent masts on ships and bunker vessels and pressure relief valves on road tankers are normally as high as possible to minimise the impact of venting on the ground and on staff. However, if bunkering takes place
© Society for Gas as a Marine Fuel 50
alongside a significantly larger vessel, the vent mast may present a hazard to the bunkering vessel.
If LNG becomes trapped in a sealed system – for example, if LNG is left in a pipe where the valves at both ends of the pipe are closed – it will start to warm as heat passes through the insulation. Some of the LNG will start to boil, increasing the pressure in the system. If the LNG-containing system or pipe work is not protected by a pressure-relieving device, the system will ultimately rupture, releasing a mixture of BOG and LNG into the atmosphere.
Venting Scenarios
Venting scenarios are outside the scope of this work but SGMF has evaluated them in outline to confirm that LNG leaks are the principal determinant of the Safety Zone. SGMF is unable to offer guidance at this time but makes the following comments:
• The IGF Code (section 6.7.2) defines the minimum distances between vents and air intakes to be at least 10 m. In the vent scenarios examined (up to 9 bar) by SGMF dispersing gas achieves LFL within this 10 m horizontal direction but not 50% of LFL. However, in the vertical direction the distance to LFL defined by the upwards momentum of the vent is larger – potentially as much as 30 m in the worst case modelled. So there should be no air intake above a vent. It is crucial to think about gas and LNG releases in three dimensions.
• The IGC Code for bunkering vessels is less proscriptive, limiting the height of the vent mast above the bunker vessel’s deck to B/3 or 6 m. This ensures safety for the bunkering vessel but the code is wholly silent on the impact of the vent on the LNG-fuelled vessel.
Vent pipe
Distance to LFL
20 m
30 m
Maximum vent plume size to LFL based on scouting calculations
© Society for Gas as a Marine Fuel 51
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
A4. LNG Leak Behaviour LNG leakage behaviour is primarily controlled by two factors:
• the pressure of the leak source
• the size of the hole from which the leak is occurring
If the leak has pressure behind it, the LNG will spray out. The higher the pressure the greater the length of the jet. The outside of the jet will mix vigorously with the air and vaporise. However, the centre of the jet will not mix with air and may be so limited in the amount of heat it can absorb that it remains as LNG. As gravity exerts its pull, droplets of LNG will fall out of the jet onto the surface below. This is how LNG is expected to behave at the storage pressures in Type C tanks (as defined by section 4.23 of the IGC Code) on board gas-fuelled ships, in road tankers, and in systems that use pumps to transfer LNG.
If there is no pressure behind the leak, the LNG will flow downwards under the force of gravity. The rate of the LNG leak will determine how the LNG behaves and how much, if any, reaches the surface below the leak source. If the LNG leak is small, the LNG will probably vaporise before it hits the ground. These small leaks will be hard to detect. Higher flow rate leaks will form pools. Pool size will be determined by how fast the pool is warmed by absorbing heat from the surface (water, steel of the ship, earth/concrete, and so on) under the pool when compared with how much LNG is entering the pool.
The rate at which LNG vaporises is primarily determined by heat transfer from the immediate surroundings. Atmospheric conditions have a secondary affect. To vaporise, LNG needs energy (heat); the rate of vaporisation depends on how quickly this heat can be supplied. Poorly conducting solid materials, such as earth or concrete, can provide considerable heat initially but after the first few seconds, once the immediate ground surface is cooled, the transfer rate will slow because of the time it takes for heat to conduct through the solid material. Metals, such as ship decks, are much better heat conductors than the ground and so can supply heat more quickly to the LNG. However, they remain relatively limited in their capacity to vaporise LNG. Steel used for shipbuilding becomes brittle and fractures at LNG temperatures. This typically results in the cracking of
© Society for Gas as a Marine Fuel 52
decks and structures. However, it is very unlikely that the structural integrity of the vessel will be affected sufficiently to cause loss.
Figure A1: Brittle fracture caused by an LNG spill
In most circumstances the spilling of LNG onto water will result in the formation of a pool on the surface, similar to its behaviour on land. For any sizeable spill the size of the pool on water will be smaller than on land, as it evaporates faster. However, if the LNG and water are able to mix quickly the resulting heat transfer can allow the LNG to boil so rapidly that the resulting gas creates a localised, high-pressure blast wave as it expands. This phenomenon is called a Rapid Phase Transition, or RPT. The phenomenon is not completely understood but all evidence to date suggest that an RPT will not extend the Safety Zone beyond that needed to deal with other scenarios.
Dry air is not a particularly good conductor of heat. The dispersion of LNG vapour (and any very small droplets of LNG within the cloud) and BOG can be significantly affected by the weather. A key factor is how rapidly they warm and become buoyant. This will be affected by:
• wind – as higher wind speeds cause more turbulence in the cloud, mixing it faster and diluting it
• atmospheric temperature
• the degree of humidity as the water vapour in the air will release latent heat as it condenses; this condensation also makes the cloud visible
(Note: White clouds of water particles do not solely result from LNG leaks. Any source of cold – for example, a bunkering hose or LNG transfer system – may cause localised condensation.)
© Society for Gas as a Marine Fuel 53
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Methane, the primary constituent of LNG or natural gas, is a small molecule, lighter than the constituents of air; the molecular weight of methane is 16, nitrogen 28 and oxygen 32. In normal conditions methane is therefore lighter (less dense) than air and will rise.
However, density is strongly affected by temperature. The lower the temperature, the greater the density. As LNG, BOG and natural gas are very cold immediately after vaporisation (say <-150°C) their density at atmospheric pressure is high: about two-to-three times greater than air at ambient temperature (of say 20°C). Therefore, around the leak, the LNG/BOG/NG will be heavier (denser) than air and will flow towards the ground, water surface or deck level. The BOG/natural gas will not be lighter than air until it has warmed to above about -110°C (–166°F). This is normally some distance from the leak source.
Figure A2: Cold methane is heavier than air (Test at Falck Training Facility, Rotterdam) © Penguin Energy Consultants Ltd
A5. Ignition Scenarios Within the gas cloud there is potentially a flammable mixture of BOG/natural gas which can be ignited if the conditions are correct. For a fire to occur there must be three components present:
• fuel (the vapour above the LNG or natural gas)
© Society for Gas as a Marine Fuel 54
• oxidant (air or oxygen mixed in the correct proportion) and
• energy (a spark or high temperature)
The flash point of liquefied methane, where the vapour generated by the liquid is sufficient to be ignited, is -187°C. Therefore in any LNG storage, transfer or operational scenario, LNG or natural gas may ignite if the fuel-air mixture is correct and there is sufficient energy to cause ignition.
Methane (natural gas) is only flammable within a certain concentration range in air. If there is too much methane the flame will not be able to get enough oxygen and will go out. If there is too little fuel again the flame will go out. These limiting conditions are called the Lower and Upper Flammable/Explosive Limits (LFL/LEL and UFL/UEL). The flammability range of methane (or LNG vapour/natural gas) in air is between 5% (LFL) and 15% (UFL).
Figure A3: Flammability limits of methane (LNG/natural gas)Oxygen in air
UFL in air 15%
LFL in air 5.3%
FLAMMABLE ZONE
Too little fuel (too lean) to burn)
Too much fuel (too rich) to burn)
Met
hane
%
Oxygen %
40.0%
35.0%
30.0%
25.0%
20.0%
15.0%
10.0%
5.0%
0.0%0.0% 20.0% 40.0% 60.0% 80.0% 100.0%
Methane can spontaneously ignite if it is heated to above its auto-ignition temperature, about 530°C (980°F). This contrasts with fuel oils, which have typical auto-ignition temperatures of around 250°C and can therefore be readily ignited by hot surfaces such as unlagged exhaust systems, resulting in engine room fires (see Table A1). Auto-ignition of any vapour is unlikely on a gas-fuelled vessel.
© Society for Gas as a Marine Fuel 55
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Table A1: A comparison of the characteristics of LNG and Marine Gas Oil (MGO).
Particulars LNG/natural gas Marine Gas Oil (MGO)
Flash point –187°C (–304.6°F)(Flammable gas) >60°C (140°F)
Flammability limits (vol% in air) 5 – 15 0.6 – 7.5
Auto-ignition temperature 530°C (986°F) 250–300°C (482–572°F)
Minimum Ignition Energy (MIE) in air (milli Joule (mJ)) 0.27 20
A spark may ignite the LNG vapour/natural gas. A spark is a highly concentrated form of energy in a very localised area. The energy of a spark required to ignite methane depends on its concentration. The minimum energy – that is, the most positive ignition gas mixture – must exceed 0.26 mJ (1 milli Joule = 1/1000th of a Joule). Spark energy increases rapidly, to about 50 times the minimum spark energy, at other methane concentrations. Ignition may not always occur (Table A2).
Table A2: Ignition probabilities
Strength Example ignition sources Ignition probability
Certain Pilot lights, fired heaters, flares 100%
Strong Hot work, electrical faults, smoking >50%
Medium Vehicles, substations, unclassified electrical equipment, engines, hot surfaces 5 – 50%
Weak
Office equipment, electrical appliances, mechanical sparks, static electricity <5%
Mobile/cell phones <0.001%
Negligible Intrinsically safe equipment 0%
© Society for Gas as a Marine Fuel 56
A6. LNG Fires
The thermal radiation given off from a fire is a threat to human health and can degrade and damage structural materials sufficiently to result in failure and potential escalation of the fire to other areas/systems. Thermal radiation is therefore a key parameter in determining safe distances from LNG/natural gas hazards.
Three types of fire are possible with LNG/natural gas systems. These fire types depend on the quality of the fuel, in effect whether the bulk of the fuel is gaseous or liquid, and secondly the pressure associated with the leak scenario. The three scenarios are described in Table A3.
Table A3: LNG fires
Fire Type Description
Pool
A fire burning over a pool of LNG. For leaks from LNG bunkering facilities that are relatively small (perhaps a few cubic metres) a bright, clean-burning flame is expected with high lu-minosity and very high thermal radiation levels. Pool fires are highly localised so will not define the distance to the edge of the Safety Zone, but may define the thermal radiation requirements of the safety distances close to the bunkering ship and infrastructure.
Jet
A fire associated with pressurised releases of gas, liquid or a mixture of the two. The behav-iour of a jet fire depends on the flow rate of the material in the jet. Liquid droplets may “rain out” of the jet onto the ground/water/deck surface, where they will burn as pool fires. The thermal radiation from a jet fire will be significantly higher than for pool fires. The thermal flux from a jet fire – and in extreme cases, the impact of cryogenic LNG droplets in the core of the jet – is likely to cause rapid damage and failure of surrounding unprotected structures and equip-ment. Again, although larger than pool fires, jet fires are localised and will not define the distance to the edge of the Safety Zone.
© Penguin Energy Consultants Ltd
© DNV GL Spadeadam Testing and Research
© Society for Gas as a Marine Fuel 57
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Flash
The combustion of a vapour cloud without generation of any significant pressure. A flash fire occurs when a cloud of flammable gas in an open area is ignited. An open area does not constrain the gas, allowing it to expand without restriction. In a flash fire the gas cloud will combust (“flash back”) from the point of ignition back to the source of the leak, combusting all the gas present in the flammable range (LFL to UFL) and then burning at the UFL boundary until all the hydrocarbon fuel in the cloud has been consumed. The fire will then become a jet fire or pool fire, depending on the leak source.
Methane has a flame speed in a completely open environment of 3-6 m/s (about 10-19 km/h or 6-13 mph) so the flame front takes little time to burn back. Therefore, a flash fire is of short time duration, a few seconds at most, which just covers the combustion of the leaked gas within the cloud. This provides little external damage but is likely to be fatal to anyone who is inside the cloud when it ignites. Gas clouds can travel significant distances prior to ignition. The exact distance will depend on weather conditions as these will determine how fast the cloud rises and disperses.
Ignition at the edge of a gas cloud starting
to burn back through the cloud to the fuel source, a
pool in this case
© Penguin Energy Consultants Ltd
© Society for Gas as a Marine Fuel 58
A7. Explosions
There is the potential for a drifting vapour cloud to enter an enclosed bunkering station, buildings, rooms or other confined areas, such as drains. If the flash fire burns into these areas, the gas can no longer expand and accelerates instead, potentially causing an explosion. Careful siting of the bunker station on the vessel and defining a bunkering area away from buildings should be considered. Forced ventilation of enclosed areas can also be used to ensure that the methane never reaches its flammable range.
Blast over-pressures from methane explosions (deflagrations) can be significant. However, the blast pressure falls rapidly with distance so explosions do not define the safety distance and are primarily concerned with distances close to or within the bunkering vessel.
Another type of explosion is known as a BLEVE (Boiling Liquid Expanding Vapour Explosion). Here, a strong heat source, normally a fire, heats up and boils LNG within a pressure vessel, such as a road tanker or IMO Type C tank. The tank pressurises and gas is released through the pressure-relief system. If the fire is intense enough, the pressure-relief valves may not be of sufficient capacity to vent gas quickly enough to control the tank pressure. Eventually the tank will rupture, creating a fireball and high-velocity missiles. Failure can be quicker if the flames impinge on an area of tank where there is gas rather than LNG, as the lack of boiling allows a faster temperature rise, weakening the steel in the immediate area.
BLEVEs are major events which can impact a wide area. However, they take time to develop particularly for insulated tanks (like LNG), allowing an emergency response to cool the vessel or to extinguish the fire and/or to evacuate the area.
© Society for Gas as a Marine Fuel 59
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
A8. Other Hazard Scenarios Asphyxiation (oxygen deficiency)Within a leaking LNG/gas vapour cloud, or within a space flooded with escaping vapour, the amount of air may be reduced and replaced by the heavier cold methane, which will disperse only slowly. Oxygen levels will reduce, impairing human mental performance. Initially, there is mental confusion (disorientation and inability to think straight), loss of muscle movement (clumsiness), emotional/mood changes and loss of sensation. This is followed by unconsciousness and death.
Cryogenic burns and impactsHuman flesh can be damaged if it comes into contact either directly with LNG or indirectly through uninsulated pipework. Brief exposures that would not affect skin on the face or hands can damage delicate tissues such as the eyes. Prolonged exposure of the skin or contact with cold surfaces can cause frostbite. The skin appears waxy yellow. There is no initial pain, but there is intense pain when frozen tissue thaws.Unprotected skin can stick to metal cooled by cryogenic liquids. The skin can then tear when pulled away. Even non-metallic materials are dangerous to touch at low temperatures. Prolonged breathing of extremely cold air may damage the lungs.
© Society for Gas as a Marine Fuel 60
Appendix B: Hazard Identification
There are many methods available to identify risks and hazards. SGMF’s bunkering guidelines (FP07-1, Appendix F) provide guidance on how to conduct a HAZID. The technique is also described in ISO 17776, ISO 18683 and IACS 146. The guidance is not reproduced here.
Several HAZID reports have been published for LNG bunkering and are available on the internet. For example:
• Hazard Identification Study (HAZID), LNG bunkering from bunker vessel in the Port of Hamburg, January 2012, RD-ER 2011.131 http://www.lngbunkering.org/sites/default/files/2012,%20GL,%20Hazard%20Identification%20Study%20(HAZID)%20LNG%20Bunkering%20from%20Bunker%20Vessel%20in%20the%20Port%20of%20Hamburg.pdf
• North European LNG Infrastructure project: http://www.dma.dk/themes/LNGinfrastructureproject/Documents/Documents/LNG_draft_FR_20111128_app_K.pdf
This appendix consists primarily of three tables of potential hazard scenarios that were discussed by the SGMF working group in a hazard assessment workshop. The workshop took place over a single day. The workshop was not a true HAZID as all options were not developed using keywords. The working group limited themselves to larger hazard scenarios that would define the Safety Zone. For example, many people will probably be injured by slips, trips and falls. However, the purpose of the Safety Zone is to minimise the potential to injure multiple individuals or to create significant levels of damage.
The workshop was performed on a sequential basis, starting with identifying and quantifying issues with the LNG supply source, followed by the LNG/vapour transfer systems, and finally the gas-receiving vessel. The main intention was to look at the larger events during bunkering, with its higher LNG/gas flows and greater magnitude in changes of other-variables such as pressure and level. Much of the analysis is also relevant to the connection, purging, cooling, purging and disconnection processes either side of bunkering. However, as the flows during these periods are much less, their impact has been ignored as the mitigations
© Society for Gas as a Marine Fuel 61
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
suggested would also apply.
The tables attached are generic and intended to be indicative rather than exhaustive. Local factors of the port and specific features of the bunker supply, transfer system and receiving vessel should be taken into account and should refine the answers.
Three tables are produced:
• Table B1 dealing with hazard scenarios impacting the LNG supply source
• Table B2 documenting the hazard scenarios impacting the LNG/gas transfer system and
• Table B3 examining hazard scenarios impacting the LNG receiver
© Society for Gas as a Marine Fuel 62
No.
Failu
re M
ode
Bunk
erin
g ty
peCa
uses
Cons
eque
nces
Miti
gatio
nD
efine
s Sa
fety
Zo
neCo
mm
ents
1
Rupt
ure
of
LNG
tank
by
ext
erna
l im
pact
Road
ta
nker
Colli
sion
.
Com
plet
e lo
ss o
f in
vent
ory
over
a
shor
t per
iod
of
time.
Mul
tiple
truc
ks
may
be
pres
ent
– es
cala
tion
pote
ntia
l.
Tank
co
nstru
ctio
n.Pr
ohib
it ve
hicl
es/c
argo
op
erat
ions
in
Safe
ty Z
one.
No
Up
to 2
5 to
nne
spilt
.Ex
trem
ely
unlik
ely
even
t – ri
sks
cont
rolle
d by
Saf
ety
Zone
and
m
onito
ring
area
.N
ot to
be
used
to d
efine
the
Safe
ty Z
one.
Shou
ld b
e co
nsid
ered
if E
xter
-na
l Zon
e us
ed.
Bunk
er
vess
elCo
llisi
on.
Com
plet
e lo
ss o
f in
vent
ory
from
on
e ta
nk o
ver a
sh
ort p
erio
d of
tim
e; m
ultip
le
tank
s pr
esen
t –
esca
latio
n po
tent
ial.
IGC
rule
s on
ta
nk lo
catio
n.Po
rt ru
les.
Vess
el tr
affic
sy
stem
.
No
Hig
her c
hanc
e an
d m
ore
ener
gy if
aw
ay fr
om q
uays
ide.
Exce
ptio
nal e
vent
.Sa
fety
dis
tanc
e no
t im
pact
ed
if M
arin
e Ex
clus
ion
Zone
fo
llow
ed.
Exte
rnal
zon
e.
Term
inal
Colli
sion
.Fa
ilure
of
unpr
essu
rised
ta
nks.
Pres
sure
-rel
ief
syst
em.
No
Not
a b
unke
ring-
spec
ific
even
t.
2Pu
nctu
re o
f LN
G ta
nkRo
ad
tank
er
Colli
sion
or
drop
ped
obje
ct.
Slow
loss
of i
n-ve
ntor
y. M
ultip
le
truck
s m
ay b
e pr
esen
t – e
sca-
latio
n po
tent
ial.
Tank
er ro
ute
to
vess
el: s
uffic
ient
sp
ace
for t
anke
r w
ith n
o tig
ht
corn
ers/
reve
rs-
ing
with
out a
ba
nksm
an a
nd
so o
n.
No
Prim
arily
a p
re-b
unke
ring
activ
ity.
Port
to p
rovi
de g
uida
nce
on
mon
itorin
g ar
ea.
Not
to b
e us
ed to
defi
ne th
e Sa
fety
Zon
e.Sh
ould
be
cons
ider
ed if
Ext
er-
nal Z
one
used
.
Tabl
e B1
: Haz
ard
scen
ario
s im
pact
ing
the
LNG
sup
ply
sour
ce
© Society for Gas as a Marine Fuel 63
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
2Pu
nctu
re o
f LN
G ta
nk
Bunk
er
vess
el
Colli
sion
or
drop
ped
obje
ct.
Slow
loss
of
inve
ntor
y.
Bunk
er v
esse
l ro
utin
g.Po
rt ru
les.
Traf
fic m
an-
agem
ent
syst
em.
No
Prim
arily
a p
re-b
unke
ring
activ
ity.
Port
to p
rovi
de g
uida
nce
on
mon
itorin
g ar
ea.
Not
to b
e us
ed to
defi
ne th
e Sa
fety
Zon
e.Sh
ould
be
cons
ider
ed if
Ext
er-
nal Z
one
used
.
Term
inal
Colli
sion
or
drop
ped
obje
ct.
Slow
loss
of
inve
ntor
y.
Ope
ratio
ns
and
mai
n-te
nanc
e pr
oced
ures
.Fe
ncin
g an
d se
curit
y.
No
Not
a b
unke
ring-
spec
ific
inci
dent
.N
ot to
be
used
to d
efine
the
Safe
ty Z
one.
Shou
ld b
e co
nsid
ered
if E
xter
-na
l Zon
e us
ed.
3
Leak
from
be
fore
or o
n fir
st is
olat
ion
valv
e
Road
ta
nker
Wea
r and
te
ar/
fatig
ue.
Corr
osio
n.Ex
tern
al
dam
age.
Smal
l lea
ks u
p to
in
vent
ory
of th
e ta
nk.
All-w
elde
d co
nstru
ctio
n ex
cept
flan
ge
on is
olat
ion
valve
ste
m.
Dist
ance
be
twee
n ta
nk a
nd fi
rst
isola
tion
valve
is
min
imise
d.In
spec
tions
and
m
aint
enan
ce
(road
tank
ers
ever
y yea
r, co
n-ta
iner
s eve
ry 2
.5
year
s, bu
nker
ve
ssel
s eve
ry
five
year
s).
No
LNG
leak
s fro
m o
ne ta
nk
impa
ctin
g an
othe
r.Le
aks
in w
elde
d an
d m
ain-
tain
ed s
yste
m (t
o st
anda
rds)
ar
e so
unl
ikel
y th
at th
ey d
o no
t ne
ed to
be
cons
ider
ed fo
r the
Sa
fety
Zon
e.N
eed
to c
onsi
der h
ow th
e dr
aina
ge s
yste
m o
n th
e qu
ay
wor
ks to
avo
id li
quid
LN
G
dam
agin
g ot
her s
truct
ures
or
the
ship
.
Bunk
er
vess
el
Term
inal
© Society for Gas as a Marine Fuel 64
No.
Failu
re M
ode
Bunk
erin
g ty
peCa
uses
Cons
eque
nces
Miti
gatio
nD
efine
s Sa
fety
Zo
neCo
mm
ents
4
Leak
afte
r firs
t is
olat
ion
valv
e bu
t bef
ore
trans
fer
syst
em
Road
ta
nker
Wea
r an
d te
ar/
fatig
ue.
Corr
osio
n.Ex
tern
al
dam
age.
Tem
pera
-tu
re c
yclin
g (li
ne c
ool-
dow
n an
d w
arm
ing)
.
Smal
l-med
ium
le
aks
of s
hort
dura
tion.
Isol
atio
n (in
clud
ing
ESD
) va
lves
redu
ce
inve
ntor
y le
aked
.ES
D v
alve
tim
-in
g to
clo
se.
ESD
val
ve
rem
otel
y op
erat
ed.
Yes
Freq
uent
, sm
all-m
ediu
m
leak
s.St
ress
es o
f war
min
g up
and
co
olin
g do
wn.
M
ust b
e co
nsid
ered
for S
afet
y Zo
ne.
Larg
er le
ak ra
tes
poss
ible
for
bunk
er v
esse
l and
term
inal
.Ro
ad ta
nker
40
mm
pip
ewor
k –
1 mm
leak
from
gas
ket
failu
re.
Bunk
er
vess
el
Term
inal
5O
verfi
ll su
pply
ta
nk
Road
ta
nker
Reve
rse
flow
.In
corr
ect
valv
e op
erat
ion/
line
out o
f m
ultip
le
tank
s.
LNG
leak
to
atm
osph
ere
from
re
lief v
alve
s.Ch
eck-
lists
.
Yes
(may
de
fine
vert
ical
lim
it)
Two
or m
ore
tank
s m
anifo
ld-
ed to
geth
er w
hich
bec
ause
of
val
ving
end
s up
with
one
bu
nker
tank
pum
ping
into
the
othe
r.Su
pply
tank
cou
ld b
e ov
erfil
led
by re
vers
e flo
w fr
om s
hip’
s ta
nk u
nder
gra
vity
or p
ress
ure
diffe
rent
ial.
Bunk
er
vess
el
Term
inal
Tabl
e B1
con
tnue
d: H
azar
d sc
enar
ios
impa
ctin
g th
e LN
G s
uppl
y so
urce
© Society for Gas as a Marine Fuel 65
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
6O
ver-
pr
essu
risat
ion
of s
uppl
y ta
nk
Road
ta
nker
Hig
h pr
essu
re in
ga
s-fu
elle
d ve
ssel
ta
nks.
Unp
ress
ur-
ised
tank
un
able
to
cope
with
va
pour
re
turn
vo
lum
e.Pr
essu
re-
relie
f too
sm
all.
Fire
.
Relie
f val
ve li
fts –
ga
s re
leas
ed.
Pote
ntia
l fai
lure
of
tank
s if
re
lief s
yste
m
inad
equa
te.
Pres
sure
- re
lief s
yste
m.
Yes
(may
de
fine
vert
ical
lim
it)
Liftin
g of
relie
f val
ves.
Gas
clo
ud a
t ven
t mas
t of
supp
ly ta
nk (e
spec
ially
mem
-br
ane/
type
A o
r B ta
nk).
Gas
-fuel
led
ship
tank
pr
essu
rises
.Ch
eck
bunk
er v
esse
l ven
t m
ast l
ocat
ion.
Fire
cas
e fro
m IG
C m
ay n
ot
repr
esen
t wor
st c
ase.
Flas
hing
gas
dur
ing
cool
-do
wn.
Bunk
er
vess
el
Term
inal
© Society for Gas as a Marine Fuel 66
No.
Failu
re M
ode
Bunk
erin
g ty
peCa
uses
Cons
eque
nces
Miti
gatio
nD
efine
s Sa
fety
Zo
neCo
mm
ents
7
Supp
ly ta
nk
mov
ing
away
du
ring
LNG
tra
nsfe
r
Road
ta
nker
Road
tank
er
driv
es o
ff or
ro
lls a
way
.Tr
ansf
er s
yste
m
over
-ext
ende
d an
d le
aks/
ru
ptur
es.
Dry-
brea
k co
uplin
g.Ai
r-bra
king
sys-
tem
eng
aged
on
ope
ning
do
ghou
se.
Choc
king
of
whe
els.
Turn
ing
off
engi
ne.
If tru
ck h
ose
perm
anen
tly
conn
ecte
d in
terlo
cks
to
brak
ing
syst
em.
No
(dry
-dis
-co
nnec
t pr
esen
t un
der
IGF
Code
)
Incl
uded
in S
afet
y Zo
ne c
al-
cula
tions
unl
ess
a dr
y br
eak
is p
rese
nt w
hen
the
scen
ario
be
com
es th
e sa
me
as m
any
othe
rs.
Bunk
er
vess
el
Bunk
er v
es-
sel m
oves
aw
ay.
Moo
ring
brea
k.
Dry-
brea
k co
nnec
tor.
ESD
syste
m.
Term
inal
Not
po
ssib
le.
n/a
n/a
n/a
n/a
Tabl
e B1
con
tnue
d: H
azar
d sc
enar
ios
impa
ctin
g th
e LN
G s
uppl
y so
urce
© Society for Gas as a Marine Fuel 67
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
8Fi
re c
lose
to
tank
Road
ta
nker
Die
sel s
pill
and
fire
in
engi
ne.
Pote
ntia
l for
fire
to
cau
se ru
ptur
e an
d BL
EVE.
Fire
fight
ing
equi
pmen
t and
tra
inin
g.Tim
e to
ev
acua
te.
Emer
genc
y re
spon
se p
lan.
No
Full
LNG
tank
s w
ould
be
very
di
fficu
lt to
BLE
VE. E
mpt
y LN
G
tank
mor
e lik
ely
to B
LEVE
.Vi
abili
ty o
f eve
nt b
eing
exp
eri-
men
tally
det
erm
ined
.A
slo
w h
azar
d no
t an
imm
edi-
ate
one
so n
ot to
be
used
for
Safe
ty Z
one.
Shou
ld b
e co
nsid
ered
if E
xter
-na
l Zon
e us
ed.
Burn
ing
rubb
ish.
Hot
wor
k.
Liftin
g of
relie
f va
lves
. Gas
re-
leas
e ve
ry li
kely
to
be
igni
ted.
Cont
rolle
d fir
e.Th
erm
al ra
diat
ion.
Mul
tiple
tank
s m
ay b
e pr
esen
t –
esca
latio
n po
tent
ial.
Safe
ty Z
one.
Set a
ver
tical
di
stan
ce to
the
Safe
ty Z
one.
Emer
genc
y re
spon
se p
lan.
No
Will
ven
t ver
tical
ly.N
o ac
tive
equi
pmen
t or p
as-
seng
ers
abov
e bu
nker
sys
tem
re
lief v
alve
s/ve
nt m
ast.
Bunk
er
vess
elFi
re
Term
inal
Fire
9Ta
nker
roll
over
Ro
ad
tank
er
Man
oeu-
vrin
g ar
ound
co
rner
s to
o fa
st w
ith a
hi
gh c
en-
tre-o
f-gra
vity
LNG
car
go.
Collis
ion
with
ano
ther
ve
hicle
.
Dam
age
to ta
nk,
valv
ing
and/
or p
ipew
ork
lead
ing
to a
LN
G
leak
.
Spee
d lim
its.
Driv
er tr
aini
ng.
No
Not
incl
uded
in S
afet
y Zo
ne a
s ve
hicl
e ne
eds
to b
e m
ovin
g at
sp
eed
whi
ch is
not
pos
sibl
e in
the
Safe
ty Z
one
as ta
nker
is
stat
iona
ry.
© Society for Gas as a Marine Fuel 68
Tabl
e B2
: Haz
ard
scen
ario
s im
pact
ing
the
LNG
tran
sfer
sys
tem
No.
Failu
re M
ode
Bunk
erin
g ty
peCa
uses
Cons
eque
nces
Miti
gatio
nD
efine
s Sa
fety
Zo
neCo
mm
ents
10Tr
ansf
er
syst
em
rupt
ure
Gen
eral
Tow
aw
ay
with
out
dry-
brea
k co
uplin
g pr
esen
t.
Hos
e, fl
ange
an
d/or
con
-ne
ctio
n br
eaks
le
adin
g to
tota
l fa
ilure
and
LN
G
leak
up
to w
hole
ta
nk c
onte
nts.
Pum
ped
syst
em
will
limit fl
ow a
s pu
mp
hits
top
of p
erfo
rman
ce
curv
e. P
ress
ure
diffe
rent
ial w
ill de
cay.
Dry-
brea
k co
uplin
g.
No
Flow
may
be
from
bot
h sh
ip
and
tank
er u
ntil
ESD
ope
ratio
n.H
oses
leak
bef
ore
rupt
ure.
Scen
ario
not
to b
e co
nsid
ered
fo
r Saf
ety
Zone
as
dry-
brea
k co
uplin
g us
ed.
Dro
pped
ob
ject
.SI
MO
Ps
proc
edur
es.
Hos
e ex
pect
ed to
sur
vive
dr
oppe
d ob
ject
.
11Tr
ansf
er
syst
em le
akG
ener
al
Wea
r and
te
ar/f
atig
ue.
Exte
rnal
da
mag
e.W
rong
de
sign
/ m
ater
ials
.D
iffer
entia
l m
ovem
ent
(bun
ker v
es-
sel l
arge
r po
tent
ial).
Smal
l hol
e in
tra
nsfe
r sys
tem
.
Insp
ectio
n of
ho
se.
Repl
acem
ent
of h
oses
.ES
D s
yste
m to
lim
it le
ak s
ize.
Yes
Cont
ents
of t
rans
fer s
yste
m
plus
flow
for r
eact
ion/
ESD
cl
osur
e.TN
O re
sear
ch s
how
s ho
se
leak
bef
ore
failu
re.
EN14
74.
Leak
bet
wee
n sh
ips
whi
ch a
ct
as d
ispe
rsio
n ba
rrie
rs.
© Society for Gas as a Marine Fuel 69
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
11Tr
ansf
er
syst
em le
ak
Ont
o st
eel
Wea
r and
te
ar/f
atig
ue.
Exte
rnal
da
mag
e.W
rong
de
sign
/ m
ater
ials
.D
iffer
entia
l m
ovem
ent
(bun
ker v
es-
sel l
arge
r po
tent
ial).
Britt
le fr
actu
re
of s
hip
/ bu
nker
ve
ssel
dec
k or
si
de s
hell.
Stru
ctur
al d
am-
age
(ver
y un
likel
y ve
ssel
sin
ks).
Wat
er s
pray
s be
neat
h M
an-
ifold
s.D
rip tr
ays
bene
ath
man
ifold
.M
anifo
ld
wat
ch/C
CTV.
ESD
sys
tem
.G
as d
etec
tion.
No
Britt
le fr
actu
re o
ver s
mal
l are
a un
der s
pill
shou
ld n
ot b
e su
ffici
ent t
o en
dang
er s
hip.
Ont
o w
ater
RPT
(Rap
id P
hase
Tr
ansi
tion)
.D
amag
e to
the
hull.
LN
G ta
nks
spec
ified
in th
e IG
F Co
de a
re lo
-ca
ted
away
from
th
e ou
ter h
ull
and
so w
ould
no
t be
dam
aged
by
an
RPT.
Drip
tray
s be
neat
h m
anifo
ld.
No
RPTs
unp
redi
ctab
le.
Initi
atin
g ev
ent m
ore
com
mon
w
ith w
aves
whi
ch a
re u
nlik
ely
to b
e pr
esen
t.Bl
ast o
ver-
pres
sure
like
ly to
da
mag
e ra
ther
than
end
an-
ger v
esse
ls.
Blas
t ove
r-pr
essu
re d
ecay
s ra
pidl
y.
© Society for Gas as a Marine Fuel 70
Tabl
e B2
con
tinue
d: H
azar
d sc
enar
ios
impa
ctin
g th
e LN
G tr
ansf
er s
yste
m
No.
Failu
re M
ode
Bunk
erin
g ty
peCa
uses
Cons
eque
nces
Miti
gatio
nD
efine
s Sa
fety
Zo
neCo
mm
ents
11Tr
ansf
er
syst
em le
ak
Qui
ck
conn
ecto
r
Poor
co
nnec
tion.
Smal
l lea
ks o
f sh
ort d
urat
ion.
Can
be
isol
ated
–
shor
t dur
atio
n.Le
ak-c
heck
ing
proc
edur
e.
Yes
Dam
age
to q
uick
-rel
ease
sy
stem
gas
ket.
Uni
t can
not f
all o
ff th
e ta
nker
to
pro
duce
med
ium
leak
. Sm
all l
eak
from
mis
alig
nmen
t po
ssib
le.
Stre
ss o
n co
uplin
g. L
NG
isol
at-
ed o
n re
mov
al –
leak
lim
ited.
Flan
ged
coup
ling
Inad
equa
te ti
ghte
ning
and
re
-tigh
teni
ng a
fter c
ool-d
own.
Mis
alig
nmen
t.D
amag
ed g
aske
t or fl
ange
.St
ress
on
conn
ectio
n.
Scre
wed
co
uplin
g
Inad
equa
te ti
ghte
ning
, or
re-ti
ghte
ning
afte
r coo
l-dow
n.M
isal
ignm
ent.
Dam
aged
or w
orn
conn
ecto
r.St
ress
on
conn
ectio
n.
© Society for Gas as a Marine Fuel 71
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
12O
ver-
pr
essu
risat
ion
Gen
eral
LNG
bl
ocke
d in
.
LNG
trap
ped
in
syst
em. P
ress
ur-
isat
ion
of s
yste
m
as L
NG
boi
ls.
Equi
pmen
t fa
ilure
. Re
leas
e of
gas
to
atm
osph
ere.
Ther
mal
relie
f va
lves
on
ship
an
d su
pply
sy
stem
.Ve
ry s
mal
l in
vent
ory
trapp
ed o
r bl
ocke
d in
.Tr
aini
ng.
Ope
ratin
g pr
oced
ures
Yes
LNG
blo
cked
in, w
arm
s an
d va
poris
es.
Hig
h pr
essu
re b
ut s
mal
l LN
G
quan
tity.
Gen
eral
Trap
ped
LNG
by
Impr
oper
/pr
emat
ure
rem
oval
of
hos
e co
nnec
tor
by fo
rce.
Hum
an
erro
r.
13Le
ak fr
om
vapo
ur re
turn
Gen
eral
Wea
r and
te
ar/f
atig
ue.
Exte
rnal
da
mag
e.W
rong
de
sign
/m
ater
ials
.D
iffer
entia
l m
ovem
ent
(bun
ker
vess
el
larg
er
pote
ntia
l).
Smal
l hol
es.
Insp
ectio
n of
ho
se.
Repl
acem
ent
of h
oses
.ES
D s
yste
m to
lim
it le
ak s
ize.
Yes
Sam
e as
LN
G le
ak b
ut lo
wer
pr
essu
re a
nd th
eref
ore
mas
s flo
w.
Leak
will
be
or b
ecom
e
buoy
ant v
ery
quic
kly.
Mor
e en
ergy
in a
gas
than
liq
uid.
© Society for Gas as a Marine Fuel 72
Tabl
e B3
: Haz
ard
scen
ario
s im
pact
ing
the
gas-
rece
ivin
g ve
ssel
No.
Failu
re M
ode
Bunk
erin
g ty
peCa
uses
Cons
eque
nces
Miti
gatio
nD
efine
s Sa
fety
Zo
neCo
mm
ents
14Ru
ptur
e of
bu
nker
ing
pipe
wor
kA
ll
Colli
sion
of
rece
ivin
g ve
ssel
with
an
othe
r ve
ssel
.
Britt
le fr
actu
re o
f sh
ip d
eck
or s
ide
shel
l.St
ruct
ural
dam
-ag
e (v
ery
unlik
ely
vess
el s
inks
).
Doub
le-w
alle
d pi
pe re
quire
d by
IGF C
ode.
No
Mar
ine
Excl
usio
n Zo
ne a
nd
Mon
itorin
g &
Sec
urity
Are
a sh
ould
pre
vent
.N
ot to
be
cons
ider
ed in
Saf
ety
Zone
cal
cula
tion
Dro
pped
ob
ject
.SI
MO
Ps
proc
edur
es.
15
Rupt
ure
of
rece
ivin
g ta
nk
by e
xter
nal
impa
ct
All
Collis
ion
of
rece
ivin
g ve
ssel
with
an
othe
r ve
ssel
.
Britt
le fr
actu
re
of s
hip
/ bu
nker
ve
ssel
dec
k or
si
de s
hell.
Stru
ctur
al d
am-
age
(ver
y un
likel
y ve
ssel
sin
ks).
IGF
Code
tank
lo
catio
n.N
o
Not
bun
kerin
g-sp
ecifi
c.M
arin
e Ex
clus
ion
Zone
and
M
onito
ring
& S
ecur
ity A
rea
shou
ld p
reve
nt.
Not
to b
e co
nsid
ered
in S
afet
y Zo
ne c
alcu
latio
n.
16
Leak
of
bunk
erin
g pi
pew
ork
betw
een
bunk
er s
tatio
n an
d ta
nk
All
Wea
r and
te
ar/f
atig
ue.
Exte
rnal
da
mag
e.D
ropp
ed
obje
ct.
Wro
ng
desi
gn/
mat
eria
l.
Britt
le fr
actu
re o
f sh
ip d
eck
or s
ide
shel
l.St
ruct
ural
dam
-ag
e (v
ery
unlik
ely
vess
el s
inks
).
Dou
ble-
wal
led
pipe
requ
ired
by IG
F Co
de.
Drip
tray
s be
neat
h m
anifo
ld.
SIM
OPs
pr
oced
ures
.
No
Pipe
in p
ipe
syst
em a
nd
purg
ing.
Haz
ardo
us Z
one
issu
e no
t Sa
fety
Zon
e.
© Society for Gas as a Marine Fuel 73
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
17
Leak
of
bunk
erin
g pi
pew
ork
in
sem
i- co
nfine
d ar
ea
(for e
xam
ple,
se
mi-
encl
osed
shi
p m
anifo
ld)
All
Wea
r and
te
ar/f
atig
ue.
Exte
rnal
da
mag
e.W
rong
de
sign
/m
ater
ials
.
Britt
le fr
actu
re o
f sh
ip d
eck
or s
ide
shel
l.
Stru
ctur
al d
am-
age
(ver
y un
likel
y ve
ssel
sin
ks).
Pote
ntia
l for
ex
plos
ion
if ig
nite
d.
Haz
ardo
us
Zone
s w
ith n
o ig
nitio
n so
urce
s.Ve
ntila
tion.
Gas
det
ectio
n.Cr
yoge
nic
prot
ectio
n (d
rip
trays
).
Yes
Sem
i-enc
lose
d bu
nker
spa
ce
shou
ld b
e un
man
ned
durin
g bu
nker
ing.
Onl
y sm
all l
eaks
pos
sibl
e du
ring
conn
ectio
n.
18O
verfi
lling
re
ceiv
ing
tank
All
Failu
re
of in
stru
-m
enta
tion
/ co
ntro
l sy
stem
.H
uman
er
ror –
fillin
g w
rong
tank
.
Pres
sure
-rel
ief
valv
es li
ft.In
ext
rem
e ca
ses
LNG
dro
plet
s ve
nted
alo
ngsi
de
gas.
Two
inde
-pe
nden
t trip
sy
stem
s un
der
IGF
Code
/Cl
ass
rule
s.LN
G in
to v
ent
syst
em a
nd
pres
sure
relie
f.
No
Mul
tiple
failu
res
requ
ired.
19O
ver-
pres
-su
risat
ion
of
rece
ivin
g ta
nkA
ll
Inco
rrec
t fil
ling.
Fire
.In
adeq
uate
or
no
va-
pour
retu
rn.
Pres
sure
- re
lief s
yste
m
unde
r-siz
ed
Pres
sure
-rel
ief
valv
es li
ft.G
as c
loud
from
ve
nt m
ast.
Vapo
ur-r
etur
n sy
stem
.Pr
essu
re-r
elie
f sy
stem
.
Yes
Nee
d to
con
side
r ven
t mas
t lo
catio
n an
d im
pact
on
Safe
ty
Zone
.
© Society for Gas as a Marine Fuel 74
Appendix C: Calculating Distances for the Safety ZoneSGMF has conducted significant research to develop a calculation methodology that allows Safety Zone dimensions to be conservatively estimated – reliably and consistently – for a wide range of bunkering configurations, flow rates and locations. This appendix describes how the methodology was developed and the data that forms its basis.
C1. Scenario Selection
The HAZID exercise reported in Appendix B is the basis of scenario selection.
A range of failure scenarios were considered, covering both LNG and vapour/BOG leaks. Other scenarios were not considered as they are considered to be mitigated by the operation of the Monitoring & Security Area in conjunction with the Safety Zone, require multiple failures, or occur over a significant timescale where operator intervention or, in extreme cases, involvement of the emergency services should prevent escalation.
Leaks are possible during bunkering from:
• the transfer system itself
• connections between the transfer system and the bunker supply/receiving vessel
• piping components such as valves and non-welded connections (flanges) in pipework and the transfer system
• instrumentation connections and fittings
The following leak scenarios are possible and need to be assessed for their impact on the safety distance:
• LNG leak from the transfer system
• gas leak from the vapour return system (if any)
• gas release from LNG supply source (for example, relief valves on a road tanker or bunker vessel)
• gas release from the gas-fuelled ship via the vent mast
© Society for Gas as a Marine Fuel 75
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
• LNG transfer system blocked in and over-pressurising
• premature disconnection (before purging) of the LNG transfer system
Leaks from the vapour return system (if used) or from the vents on the gas-fuelled vessel (vent mast) or LNG supply infrastructure are all at relatively low pressure. They therefore have little momentum behind them. As there is limited need for energy from the atmosphere (as there is no vaporisation), they disperse in relatively short distances compared with LNG leaks. Gas clouds produced from LNG leaks therefore define the extent of the Safety Zone.
A variety of consequences can occur from leaks, including:
• flash fires
• jet fires
• pool fires
• explosions
• Rapid Phase Transitions (RPTs)
• structural damage resulting from cryogenic spills
Individuals may be harmed – or potentially killed – and property will potentially be damaged by all of these scenarios. However, most have a limited impact area and therefore do not define the size of the Safety Zone. For example, pool fires and jet fires are highly localised and so will not define the distance to the edge of the Safety Zone but may define the thermal radiation survival requirements close to the bunkering ship and infrastructure.
Gas clouds formed by leaking LNG or cold boil-off gas/vapour return are initially heavier than air and can travel significant distances before igniting. The Safety Zone is defined by the maximum extent of a LNG spill, evaporation of an LNG pool, or gas release which can subsequently be ignited. On this basis the potential of a delayed ignition of a gas cloud to cause a flash fire is the event that defines the safety distance.
© Society for Gas as a Marine Fuel 76
C2. Leak Behaviour Many factors determine how far a gas cloud will spread and remain within flammable concentration limits. SGMF employed expert consultants to review a range of scenarios, including those covered by the HAZID, to investigate the effect of different parameters on the distance to the edge of the flammable region (LFL and 50% LFL) of a dispersing gas cloud. The flammable limit is normally within the visible cloud. This detailed analysis has been used in the formulation of SGMF’s calculation methods for determining the safety distance.
A competitive tender was issued to four expert consultancies and subsequently awarded to the UK hazards group within DNV GL. One particular attraction of the DNV GL bid was its continuity through its British Gas R&D heritage of operating the Spadeadam test site in northern England. This is where most of the LNG hazards research that underpins current LNG models was undertaken. The modelling was performed phenomenologically using models calibrated against Spadeadam test data. This allowed rapid calculations of multiple parameters and consideration of three-dimensional events which avoided some of the well-known limitations of some software and the expense and time of developing geometry-specific Computational Fluid Dynamic (CFD) models.
LNG leaking from a transfer system or any other LNG-containing equipment is extremely cold compared with the local environment and vaporises into natural gas as rapidly as heat can be absorbed from its surroundings. How fast is determined by the quality of the LNG and its pressure. In most conditions the LNG exits the leaking hole at an elevated pressure.
Gas cloud behaviour consists of three stages that define how a cloud disperses. This starts with momentum, then buoyancy and finally atmospheric effects. These are summarised in Figures C1 and C2 and Table C1.
© Society for Gas as a Marine Fuel 77
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Figure C1: Dispersion Stages
Figure C2: Experimental gas dispersion test at DNV GL’s Spadeadam test site
© DNV GL Spadeadam Testing and Research
Table C1: Stages of dispersion
Stage Behaviour
Momentum
Entrainment of air (mixing) driven by differences in velocity between the jet and the air. There is less entrainment for small differences in velocity (high wind speed, co-flowing wind) and more entrainment for large differ-ences in velocity (low wind speeds, cross winds).
Buoyancy
Buoyancy can influence the trajectory of the jet. Elevated denser-than-air jets fall towards the ground, where entrainment on the bottom edge of the jet is inhibited. Denser-than-air plumes in contact with the ground “slump” outwards, increasing the entrainment on the edges of the plume. However, the density difference of lighter air above the heavier plume inhibits mixing on the top of the plume.
Atmospheric effects
Entrainment driven by atmospheric turbulence, with more entrainment for higher wind speeds, unstable atmospheres and larger surface rough-ness (size of obstacles that cause or increase turbulence along the bot-tom surface of the plume), and less entrainment for lower wind speeds, stable atmospheres and smaller surface roughness.
© Society for Gas as a Marine Fuel 78
Depending on the nature of the release and the concentrations being considered, some of these effects may dominate and some may not be relevant. For example, a release of LNG from a source at more than 5 barg would have a velocity exceeding 50 m/s. This would usually be more than six times the wind speed, so the release could be considered to have high momentum, and to behave like a jet. If flammability is the main concern, then the dispersion may only be of interest for concentrations exceeding a typical LFL of around 5% of vapour by volume. The dispersing plume might still be travelling significantly faster than the wind when this concentration is reached, so the atmospheric effects may be small.
The factors which determine how far a gas cloud will spread and remain within flammable concentration limits include:
• LNG transfer flow rate, temperature and pressure
• hole size
• different orientations of leaks: vertically, horizontally and downwards
• various climatic conditions around wind speed, climatic stability, ambient temperature and humidity
• a range of LNG compositions and physical properties
• different geometries/topographies for releases over land and sea and at different elevations
• various durations of release (depending on the type of ESD system)
A parametric analysis was performed to determine how each variable contributed to the gas dispersion range and therefore the safety distance. An example of one such analysis is shown in Figure C3.
Figure C3: Example of one of the parametric analyses
© Society for Gas as a Marine Fuel 79
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Importantly, these parameters interact with each other which means that some effects must be considered together. This increases the complexity of the model, resulting in the need to consider about 1.4 million data points for eight parameters. These parameters are described below:
Leak OrientationThree leak orientations define LNG behaviour: vertically upwards, horizontal and vertically downwards. Leaks at all angles in between can be shown to travel less far than these main orientations.
For pressurised releases of LNG all three stages can be important. In particular, different orientations of the release can behave in different ways:
• vertically upward releases initially flow upwards before eventually stalling and returning to the water/land surface beneath; hence, the vertical extent of the hazard is sensitive to parameters which affect the stalling of the plume, such as buoyancy and the initial velocity of the release.
• horizontal releases maintain their horizontal momentum, with density effects causing the plume to drop towards the ground or water surface.
• downward releases impacting on the ground or the water surface can lose most of their momentum and form a pool on the surface; the vapour evaporating from this pool forms a low-momentum plume which is less dense than the initial two-phase jet. The dispersion of this plume can be sensitive to the atmospheric conditions, but the momentum of the release is relatively less important.
These behaviours are described in more detail in the following table and in Figures C4, C5 and C6:
© Society for Gas as a Marine Fuel 80
Horizontal JetsBased on 49 large-scale experiments conducted at Spadeadam at low pressure for an international joint industry project in 1998, the following behaviour is expected:• the dispersing LNG vapour is visible
because the cold LNG causes water in the atmosphere to condense, forming a white fog.
• liquid LNG is probably present only in the initial stages of the jet.
• the jet initially travels in a horizontal di-rection, then drops slowly down towards ground level
• there was little if any rain-out of liquid LNG in the horizontal experiments, although there was an ice patch where the release landed on the ground
• after landing on the ground the low-pressure releases, which were at pressures that could be used in bunkering operations, formed low-lying clouds which were affected by the wind direction and speed.
• the width of the clouds increases rapidly after the release has landed on the ground, and the height of the cloud is very much less than its width
Vertically Upwards JetsWhen a release of LNG is directed vertically upwards, the jet initially travels upwards before stalling and falling back towards the ground. Based on the same Spadeadam interna-tional joint industry project, the following behaviour is expected:• the vertical jets are tipped over in the
direction of the wind. • despite the initial upward momentum of
the release, the vertical jets eventually stall and fall back towards the ground.
• in low wind speeds the effects of gravity dominates, and the flow sinks to the ground to form a dispersing plume at low level
Figure C4: Horizontal jet experiment © DNV GL Spadeadam Testing and Research
Figure C5: Vertically upwards jet experiment
© DNV GL Spadeadam Testing and Research
© Society for Gas as a Marine Fuel 81
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Vertically Downwards JetsThere have been several international experimental programmes investigating the be-haviour of the LNG pools created wholly on water (such as Maplin Sands) and onto water but spreading over land (such as the Burro and Coyote tests). In these experiments the LNG is released in a downward direction. The releases are gen-erally close to the ground, and the release systems are designed to give a low momen-tum source – for example, by spraying LNG out of several sources over the water pool. This is slightly different from the accidental releases which may occur during bunkering operations, where the source could be higher above the ground, and might lose less of its momentum.Vertically downward releases are assumed to impact on the ground or water surface. All the released liquid is assumed to lose its momentum, rain out, and form a pool on the surface below the release point. The liquid spreads on the warm surface and a low momentum vapour cloud evolves from the pool. Any flashed vapour from the release is also assumed to lose its momentum, and it is assumed to combine with vapour that is predicted to evolve from the pool to form a low-momentum dispersing cloud.The following behaviour is expected:• the dispersing plume is visible because it causes water vapour in the atmosphere to
condense • the source evolved vapour has a very low velocity and it disperses in the direction of
the wind • the cloud “slumps” outwards from the source because it is denser than the surround-
ing air; in low wind speeds this can cause the LNG vapour to spread upwind of the source
• the release forms a wide plume, with the height of the plume being very much less than its width
Other Leak OrientationsReleases with other orientations could also occur, as shown in Figure C6. However, the vertically upwards release (purple) gives the maximum vertical extent and the hori-zontal release (green) gives the maximum horizontal extent so these represent the most conservative options.LNG falling from a leak or raining out of a gas jet is likely to form a pool on any flat surface if it does not vaporise before it lands. The pool will continue to grow until the leak is isolated or an equilibrium is reached in which LNG evaporates from the pool at the same rate that it spills into the pool. The LNG evaporating from the pool has low/no momentum so these releases travel in the direction of and approximately at the speed of any local wind.
Figure C6: Release orientation effects
© Society for Gas as a Marine Fuel 82
LNG Transfer Flow Rate, Temperature and PressureThe larger the transfer rate, the larger the transfer equipment, the larger the potential leak and therefore the larger the safety distance.
Transfer pressure and temperature define, along with hole size, the initial momentum and so play an important role for horizontal and vertically upwards distance calculations – significantly modifying those based on flow rate alone.
The temperature and pressure of the LNG in storage prior to transfer (that is, before pumping to transfer pressure) primarily define the energy in the LNG and how much of the LNG vaporises (“flashes”) at the leak point and how much remains as a liquid. The liquid LNG vaporises as it absorbs energy from the surrounding atmosphere.
The dispersion distances are larger for higher pressures and lower temperatures, both of which give higher mass flow rates, the latter because the LNG is less likely to flash inside the hole. If the transfer pressure (in gauge) is 50% or more greater than the saturation pressure, the dispersion distance is relatively small and insensitive to the transfer pressure. If the transfer pressure is only slightly above the saturation pressure, the release continues at a similar rate after ESD, resulting in a larger dispersion distance.
For horizontal releases the results show a significant variation with the transfer temperature and pressure, with larger changes for lower pressures and higher temperatures, particularly for pressures below 10 barg, where transfers are more likely to occur.
For vertically upwards releases the stall height increases with the pressure in the transfer system but is relatively insensitive to temperature. Density of the LNG/vapour is the key parameter but this is related more to vapour fraction and flashing (pressure change) on release. Interpolation is difficult as points can be both in the vapour and liquid phases.
LNG Compositions and Physical PropertiesThe composition of the LNG has a minor impact on safety distances,
© Society for Gas as a Marine Fuel 83
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
primarily through its effect on the calculation of flammability limits (LFL and UFL) and on the level of initial flashing as the LNG exits the hole.
The composition of the LNG changes the vapour density, which can impact the distance to LFL for horizontal releases by up to about 20%. The leaner (more methane-like) the LNG, the smaller the dispersion distance. The variance of LNG saturation temperature with pressure depends on the LNG composition.
Four LNG compositions (of the 23 bulk-traded LNGs) have been included to cover this range of variation. The LNGs have been selected based on broad ranges of Calorific Value (and Methane Number) but also represent large volume-traded LNGs. The four compositions chosen for the parametric analysis and their major properties are shown in Table C2:
Table C2: LNG composition impacts
LNG source Quality Methane
contentMolecular Weight
Net Calorific
Value
Methane Number LFL
Trinidad & Tobago Lean 96.8% 16.56 35.0 88.7 4.77%
Algeria (Skikda) 91.4% 17.33 36.1 80.6 4.63%
Qatar 90.9% 17.75 37.1 74.7 4.52%
Australia (North West Shelf)
Rich 87.3% 18.56 38.8 68.8 4.34%
Notes: LNG compositions taken from the GIIGNL Annual Report for 2017 and are loaded compositions; Methane Number calculated by the AVL method; Net Calorific Value calculated at 15°C/15°C and 1.01325 bar.
Hole SizeBecause the safety distance is based on the potential hazard from a release of LNG, it needs to take into account the amount of LNG released and how quickly this happens. A key element of this is the size of the “hole” that results in the leak.
© Society for Gas as a Marine Fuel 84
Some means of selecting a representative hole size to determine the safety distance needs to be developed. Selecting a suitable hole size is a complex process and is therefore covered separately in Appendix D.
Atmospheric Conditions Atmospheric conditions can have a significant impact on gas dispersion distances. Therefore, to be effective, any analysis must include climatic factors based on the bunkering location. Temperature and humidity are primarily a function of latitude but analysis showed significant scatter unless a longitude component was also included. Temperature and relative humidity data were obtained for a number of ports worldwide. The locations of these ports is shown in Figure C7. The local variation in the values is not significant and so all locations can be modelled to sufficient accuracy using the nearest available port data point.
Figure C7: Ports used for location information
The maximum temperature used is the highest average temperature in a month recorded over several years of weather measurements. Relative humidity is taken in the evening where possible (or otherwise as a daily average). In general, the relative humidity decreases as the temperature increases, although there is more scatter in the data for locations with higher maximum temperatures.
The stability of an atmosphere is a measure of the atmosphere’s tendency to encourage or deter vertical motion. Stable atmospheric conditions and
© Society for Gas as a Marine Fuel 85
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
low wind speeds (a Pasquill stability factor of F, a stable atmosphere at night, and wind speeds of 3 m/s or less, see Table C3) lead to reduced turbulence and therefore the largest safety distances.
A Pasquill stability factor of F occurs only rarely. Its use for the External Zone is justified but for the Safety Zone a D stability factor, the most conservative daytime option, has been assumed. The difference in the two values is about 30%.
Atmospheric stability and wind speed were analysed for both horizontal and vertical releases of LNG.
Table C3: Pasquill atmospheric stability factors
Wind speed Daytime solar radiation Night-time cloud
cover
m/s Strong>700 W/m2
Moderate350 – 700 W/m2
Slight<350 W/m2 >50% <50%
<2 A A – B B E F
2 – 3 A – B B C E F
3 – 5 B B – C C D E
5 – 6 C C – D D D D
>6 C D D D D
A = very unstable B = unstable C = slightly
unstable D = neutralE =
slightly stable
F = stable
For both horizontally and vertically upwards jets, the maximum dispersion distance occurs for a low wind speed in a stable atmosphere, as wind speeds are too low to cause sufficient turbulence to disturb the temperature gradients in the atmosphere.
For vertically downwards jets, the maximum dispersion distance also occurs for a low wind speed in a stable atmosphere. The predictions are much more sensitive to the atmospheric stability than the jetted releases. There is no momentum-dominated phase, and the buoyancy and atmospheric
© Society for Gas as a Marine Fuel 86
stages of the dispersion are therefore more important. The distances to the LFL are greater than for the jet releases. The pool will usually form on the land or sea surface below the release, and, because of the low momentum of the release, the vapour tends to remain close to ground level so the vertically downward releases lead to larger safety distances.
All dispersion distances tend to increase as temperature and humidity increase. The ambient temperature is linked to the atmospheric stability, with stable atmospheres being less likely at higher ambient temperatures.
Various Durations of ReleaseThe duration of a release depends on how long it takes to detect and isolate the leak. If the hole size is small compared with the hose area, the release rate will be less than the transfer rate and therefore there will not be any pressure drop in the hose – so the release would either be detected by an operator or by a gas-detection system. Gas-detection systems are unlikely to detect the release until a flammable cloud of a reasonable size has already formed.
The IGF Code requires an emergency shut-down system to be used during bunkering. The wording in the IGF Code (Section 8.5.7) is a little ambiguous but is interpreted by many as requiring a fully automatic ESD system. In reality this may not be achievable, as some bunkering facilities – for example, older road tankers, ISO containers and some shore terminals with distant tanks – may not be fully automated, that is, semi-automatic.
SGMF has included four ESD scenarios in the model:
• fast acting fully automated ESD (requiring no human interaction) which stops the transfer in 10 seconds
• a fully automated ESD system (requiring no human interaction) which acts in 30 seconds (from leak detection to full valve closure)
• a semi-automatic ESD system which requires some level of human interaction – for example, responding to an alarm and hitting the ESD button or manually closing a valve – and therefore takes longer to implement; two minutes has been selected as this is a common value in oil, gas and chemicals systems where similar activities take place
© Society for Gas as a Marine Fuel 87
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
• a fully manual ESD system which requires multiple human actions to occur, where a ten-minute period is commonly employed
To be effective, the ESD system must act faster than a LNG/gas leak achieves its maximum extent.
ESD actions include several steps which contribute to the overall time for the release to be isolated. These include:
• The time it takes for the release to be detected – The leak has to be detected by an instrument and, in many cases , two or more to avoid spurious ESDs. The detectors must be correctly specified and sited, and any weather/dust limitations recognised. Detection will not be instantaneous. Small leaks take much longer to detect than large leaks.
• The time it takes for the ESD system to respond – The signals from the detector must reach the control system, which must then analyse the data and react, either by flagging an alarm for human action or automatically. If the latter, this process is fast.
• The time it takes for the valves to close – The valve must physically be closed either by an actuator or by a human. The bigger the valve, the slower the closure time. Closure times may need to be extended to stop over-pressures (surges) resulting from fast-acting valves.
The time to isolation needs to be significantly less than the times shown in Table C4 for isolation through ESD to reduce the maximum dispersion distance for horizontal and vertical free jet releases.
Table C4: Time for horizontal jets to disperse
Hole size Time to disperse to LFL
5 mm 2-5 s
10mm 5-12 s
20 mm 10-27 s
© Society for Gas as a Marine Fuel 88
ESD type and action time are unlikely to reduce the maximum dispersion distance for horizontal free jet releases.
For a vertically upwards jet, the relevant timescale is the time it takes for the jet to reach its maximum height and to start returning to ground level. This will take less time than a horizontal jet will take to disperse below the LFL, so vertical jets are less likely to be affected by isolation than horizontal jets.
The vertically downward releases are sensitive to the duration of the release, which depends on the time taken for the hose to be isolated. This is because the time to isolate the release influences the total mass of LNG which flows into the pool.
Figure C8 shows how a LNG liquid pool onto the ground behaves before and after isolation. In this case, isolation occurs at 30 seconds. LNG (blue line) continues to flow into the pool as the system drains until about 220 seconds. Evaporation from the pool (yellow) dominates the gas mass flow and is 7-8 times larger than gas flashing from the original LNG leak (red).
Risavika incident learning points
A break-away coupling became overstressed and leaked.
The gas detector was located above the manifold but most of the leak was heavier than air so the gas detector did not trigger immediately.
The hose watch also had other tasks to perform and so was not continuously monitoring the CCTV.
Detection took 42 seconds. Over 140 kg of LNG leaked.
© Society for Gas as a Marine Fuel 89
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Figure C8: LNG leak into a pool and vaporisation
Different Geometries/Topographies for Releases Over Land and Sea and at Different ElevationsRelease elevation has a significant effect for horizontal jets. If a hose is on the ground, air cannot get beneath the dispersing gas cloud and dispersion is considerably reduced, leading to larger safety distances. The effect is very specific; even minor changes in elevation to perhaps a meter make considerable differences, but beyond this the effect is minor.
The LNG or cold vapour behaves differently depending on whether it impacts water, soil/concrete or the steel hull of a vessel.
The surface roughness has more effect on the later, atmospheric, stages of the dispersion and less effect on the initial high-momentum stage. Horizontal releases are free jets, that is, they are not assumed to directly impact any large obstacles, although they will fall gradually downwards due to gravity, and may land on the ground. With no impact to slow the jet, it generally continues with a high velocity and the surface roughness would be expected to have a small effect, although a greater effect would be expected at larger distances as the jet gradually slows down.
© Society for Gas as a Marine Fuel 90
Decks of ships and land areas such as jetties are likely to include a range of obstacles of varying sizes which will increase turbulence along the bottom surface of a dispersing gas plume. This affects the amount of mixing between the gas and the surrounding air. In contrast, water is flat (or has minor waves) and so has a very low surface roughness (about 10% of land/deck scenarios).
The surface on which the LNG spreads will also affect the evaporation rate. Warmer surfaces will generally give greater heat transfer to the pool, leading to more rapid boiling. On this basis LNG spilt onto water will evaporate faster than the same amount of LNG spilt onto earth, concrete or steel, leading to a narrower but taller gas cloud. As this slumps it tends to dilute more than a wider, shorter cloud, and the additional dilution partially offsets the higher evaporation rate. The larger dispersion distances tend to occur for low wind speeds as the dispersing gas remains over the pool for a significant residence time – so the dispersion is affected by the evaporation rate over a period of time, and not simply the maximum evaporation rate.
Solid surfaces have a limited ability to provide heat after an initial period during which they are cooled. Heat transfer in liquids – water in this case – can continue to get significant heat transfer from neighbouring areas, so can continue higher rates of evaporation for longer.
© Society for Gas as a Marine Fuel 91
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Appendix D: Hole Sizes in LNG Transfer SystemsA variety of transfer options are available for LNG bunkering. The most frequently used are flexible hoses, but articulated hard arms are also in service. The LNG transfer system also includes valves, both manual and actuated, flanges where valves connect with pipework, and hoses and connections for instrumentation.
Figure D1: Components of an LNG transfer system
ESD ESD
LNG
ESD ESDBOG
E E E E EE E E E E
ESD Link
Vent mast
Local vent
Supplytank
Receivingtank
Dry-breakawayconnectors
Dryconnectors
Leaks can occur from all these components. A “hole size” is defined to represent these leaks as if they were circular holes. These holes are then used to calculate the leak rate used for the gas dispersion calculations. In reality some hoses exhibit small tears rather than holes.
Flanges and valves are used throughout the oil, gas and chemicals industries and there is a large body of information available on anticipated leak sizes and frequencies. With interpretation, this failure information – for example, the Process Leak for Offshore installation Frequency Assessment Model (PLOFAM) database, based on leaks in the UK and Norwegian sectors of the North Sea – can be applied to LNG bunkering.
Articulated hard arm-type LNG loading systems are assembled from multiple components, including pipes, swivels, valves, flanges and so on. Each of these components can fail but failure modes which are similar to those in piping systems are better understood. Gas industry information and operating experience is available, particularly from offshore gas production, which can be used to quantify these failures and therefore form the basis of the Safety Zone calculations.
© Society for Gas as a Marine Fuel 92
Analysis of hose-based transfer system failure is more difficult. Anecdotal comments suggest that rupture (guillotine failure) is unheard of in the LNG business and that leaks from hoses are unusual. Providing evidence to support these comments is difficult.
Information on hose failures has been reported but, on further examination, the sources and pedigrees of this data are very limited – perhaps totally inappropriate for LNG transfer. For example, LNG transfer hose failures have been based on:
• crude oil (rubber) hoses (in studies for the Dutch external safety regulatory agency, commonly known as BEVI/REVI)
• metal hoses (from nuclear and washing machine hose data from the 1970s)
• UK Health & Safety Executive failure rates for metallic hoses transferring chlorine (based on engineering judgement rather than actual failure information)
Rubber hoses are normally used for oil bunkering and in other liquefied gas industries, such as LPG and ammonia. Rubber hoses have “ruptured” causing significant loss of product and, where toxic products have been involved, fatalities. Rubbers become brittle when exposed to cryogenic conditions and so are not suitable for LNG bunkering. The only possible use of rubber in LNG bunkering would be protection of the outer layer of a cryogenic hose from environmental effects should it be designed to float or be partially or totally submerged in water.
Only two types of hose are suitable for LNG bunkering:
• composite hoses consisting of multiple layers of leak-resistant films and strength-developing fabric protected by two metal wires, one internal, one external, with metallic flange connections – described as Type 2 in the European Standard EN 13766
• metallic hoses consisting of stainless steel bellows with one or two layers of external stainless steel braiding, with threaded couplings on the ends
© Society for Gas as a Marine Fuel 93
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Requests for information on hose failures from major hose users, LNG road tanker operators, cryogenic gas companies and associations, and hose manufacturers have resulted in little concrete information. All agree that hose failure is unlikely.
A recent US research paper (“Statistical Review and Gap Analysis of LNG Failure Rate Table”, GTI project 21873, Final report, January 2017) concluded that guillotine failure was not credible but did not elaborate on what hole sizes should be used for hose failures. SGMF’s review has produced similar findings.
Data from LNG importers group, GIIGNL, in 2008 suggested that there were about 190,000 LNG truck loadings per year. This will have increased and is probably now in excess of 400,000 transfers per year. On this basis it can be credibly suggested that about five million LNG truck transfers have been performed worldwide.
Anecdotal comments from both the bulk LNG industry and LNG use at smaller scale are that there have been no guillotine failures and that leaks from transfer system failures are generally small. Data collected by SGMF on a confidential basis from 11 sites to support this position is given in Table D1. These transfers typically use 1 inch, 2 inch and 3 inch metallic hoses as these are lighter and easier to handle than composite hoses.
Leaks occur frequently during connection of the LNG transfer system. Most of these occur because of poor alignment of fittings or inadequate tightening of bolts/screw threads on flanges that have shrunk during cool-down. Both scenarios are readily detectable and should be mitigated quickly during the normal cool-down and leak-checking process. The uptake of quick connector-type devices on smaller hose systems (<6 inch) which have only limited connection options will improve initial leak performance.
Anecdotal comments from transporters of liquefied gases (liquid oxygen and nitrogen), including the British Compressed Gas Association (the UK trade body), are that no one could recall a catastrophic hose or coupling rupture. This industry has performed tens of thousands of hose transfers per year for several decades. They observe that worn couplings first
© Society for Gas as a Marine Fuel 94
Table D1: LNG hose use/failure data
Source Data provided
1 Replaced metallic hoses annually although no leakage had been detected. Over this time each hose would have undertaken about 4,000 LNG transfers.
2Replaced metallic hoses annually. Minor leakage was normally detected at about this time. Each hose would have undertaken about 3,000 LNG trans-fers.
3 Replaced metallic hoses annually although no leakage had been detected.
4Replaced metallic hoses between 5-6 months but if any exceeded this a leak-age test (at 29 bar vs hose design of 23 bar) was conducted. Small holes had been found in some hoses. This data covered over 200,000 truck loadings.
5 No hose ruptures or major leaks from 50,000 LNG transfers.
6 No serious failures from 50,000 LNG transfers.
7Hoses used for 250 LNG transfers only. Small leaks from operator errors in attaching the hose. No catastrophic hose or fitting leaks. If the hose achieves 2.5 years of life, it is replaced.
8Two incidents reported from nearly 200,000 transfers for road trucks: one was a drive-away by an LNG-fuelled truck with the dispenser hose still attached; and the second was an operator removing a quick-connect fitting prematurely.
9Hose-based LNG transfers performed for over 20 years (about 175,000 trans-fers) without a single rupture (>25 mm hole) or medium-sized (7-25 mm) hole recorded.
10 Maximum spill size over 300,000 road tanker loadings was less than 20 litres (0.02 m3).
11 230,000 truck loadings without incident.
Notes: The difference between Site 1 and 2 was that Site 1 always loads via the rear manifold of the road tanker whereas Site 2 is able to both rear and side load. Side loading is the norm, which means that the hose specified for rear loading is too long and so left coiled beside the LNG supply manifold. The tightness of the coil probably leads to greater fatigue. BASiL calculates that the spill recorded for Site 10 would be equivalent to a 5 mm hole (at a 3 bar transfer pressure and 30 s ESD time).
© Society for Gas as a Marine Fuel 95
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
begin to be difficult to seal and then leak despite any amount of effort to make them leak-tight. For road tankers, hose connection – that is, the number of transfers – is perceived to be much more important than aggregate transfer time for hose failures.
D1. Failure of Cryogenic Hoses Cryogenic hose failures do happen as evidenced in Table D1. No details of how or why the hoses failed was either recorded or provided. Table D2 shows possible failure mechanisms. Table D2 also provides, where available, information on how a cryogenic hose is expected to behave prior to failure. In most cases damage is obvious and should be found during hose inspection prior to bunkering and the hose rejected as unsuitable.
For road tankers, coupling or hose leaks/failures are most likely to occur during the first few minutes of operation, for the following reasons:
i. thermal fatigue due to cool-down
ii. hose bending during connection that causes damage due to stress
iii. damage done while pulling the hose out from storage and movement to install
iv. error in coupling installation that leads to a leak
For marine bunkering, where ship movement may stress the hose continuously, the hose failure window for item ii above will last throughout the LNG transfer.
The LNG ship-to-ship (STS) transfer business has been removing (8 inch) hoses from service for pressure testing to bursting. The results of this programme have been very positive but to date insufficient operating hours (perhaps 2000 transfers of 16 hours each through 8 hoses simultaneously) have been accrued for the results to be statistically significant.
© Society for Gas as a Marine Fuel 96
Failure mode Evidence of behaviour
Excessive flow velocity
Erosion of the metal corrugations/fabric layers
Velocity limits (12 m/s) are normally provided. High flow velocities nor-mally lead to high pressure drops so volumetric flows do not improve proportionally.
Excessive pressure
Maximum, pulsing (vibration) and shocks/spikes
The operating pressure of the hose is required to be considerably less than the design pressure. Burst testing of hoses has required pressures many times in excess of operating pressure to cause an immediate failure, even if the hose is damaged.
VacuumCreation of a vacuum can destroy the internal structure of a hose. Pos-sible causes would be draining a long LNG hose held vertically under gravity without providing a gas/nitrogen supply to maintain pressure.
Excessive motion
Extension
Over-tensioning (break-away)The Risavika LNG spill in Norway resulted from an over-tensioning of the dry-break/break-away coupling. The dry-break coupling was not de-signed correctly – it was designed for over-tensioning in the longitudinal direction whereas the strain was in the lateral direction.
In the road truck fuelling business “drive-aways” (fuelled vehicle moves off with fuelling hose still attached) have happened on several occasions. An-ecdotal comments about the use of metallic hoses in other cryogenic indus-tries suggest that the hose is actually stronger than the manifold pipework and deformation/damage will occur in the latter before the hose.
The bulk LNG industry also saw incidents in the 1970s.
The purpose of the dry-break/break-away coupling (required by the IGF Code) is to be a weak point in the connection which splits, sealing in the vast majority of the LNG, when set tensions are exceeded.
Fatigue through excessive bendingBoth metal and composite hose types fail predominantly through gradual degradation due to excessive motion (fatigue). Fatigue mostly occurs through continuous movement, mostly bending. Over-bending and multiple bending/twisting events appear to cause the most dam-age and reduction in hose life.
Bend limiters provide a degree of protection to the hose by preventing over-bending, which significantly reduces hose life.
Table D2: Cryogenic hose failure mechanisms
© Society for Gas as a Marine Fuel 97
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Failure mode Evidence of behaviour
Compression Fatigue through excessive bendingSee entry above
Torsion Frequency and magnitudeNo information found.
Vibration Fatigue through continuous movementNo information found.
External Damage
Impact by a sharp object
The only experimental work that SGMF is aware of on the impact of sharp objects has been on large (>6 inch) composite hoses for the bulk transfer of LNG. Two scenarios appear to have been examined, firstly dropping an object such as a flange (limited weight and drop height) which damages the outside of the hose but does not break the external protecting wire, and secondly failure of the external wire. In both cases, the limited data available suggests that no leak will occur from the hose at operating pressure. Burst pressure is significantly reduced but is still multiple times the operating pressure. Deformation and breakage of the supporting wire leads to the largest reduction in burst pressure but this was still several times operating pressure. The damage is visually obvious, allowing the hose to be removed from service.
Crushing or impact by a blunt object
Crushing of both metallic and composite hoses by a heavy weight, such as the LNG road tanker, has undergone very limited scientific examina-tion. Damage to the hose was clearly visible and as long as the hose had external protection from wires no leakage was found. Burst testing sug-gested that hose failure still occurred at about ten times operating pres-sure. Failure is normally around the connection between the hose and the flange/coupler rather than from the hose itself. However, in severe crushing experiments local hose failure is possible. As the damage was visible, operating procedures should ensure that the hose is not used.
Failure mechanisms for crushing and impact from a large object without sharp edges are theoretically similar. It is therefore hypothesised that hoses would perform in a similar way to crushing. One experiment on a composite hose appears to support this position.
Abrasion
One hose component against anotherNo information found.
Hose against a surface/support point.External damage to the hose can be caused by dragging the hose along the ground, resulting in abrasion. The hose should be lifted whenever possible and preferably be supported so that it is off the ground
Table D2 continued: Cryogenic hose failure mechanisms
© Society for Gas as a Marine Fuel 98
Failure mode Evidence of behaviour
External Damage (continued)
Corrosion
Operation in a marine environment is very likely to cause corrosion to the outside of the hose. Of particular concern would be loss of integrity of the external wire that helps to maintain the pressure performance and leak tightness of the hose. Corrosion of metals is normally associated with discolouration so external damage should be visible prior to use. If an outside layer is used to protect the wire, then damage to this layer, perhaps at the connection point, could lead to corrosion that would not be visible (CUI – corrosion under insulation). The experimental work from impact damage suggests that the hose can survive a few operating cycles with a failed wire.
Corrosion should be preventable by good hose handling and storage procedures and by limiting the life of the hose
Internal Damage
Corrosion
LNG is an extremely clean fluid which has had all the compounds with the highest corrosion potential removed during liquefaction. Internal cor-rosion of LNG equipment is virtually unknown. Corrosion impacts should therefore be limited to the outside of the hose.
Correct handling procedures should prevent moisture and salt entering the hose when it is not in use. It is good practice to fit end caps to the flange faces and to store hoses in a non-stressed (straight) manner on dedicated racks.
Table D2 continued: Cryogenic hose failure mechanisms
© Society for Gas as a Marine Fuel 99
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
D2. Hole Size Selection There is insufficient experience or experimental work to confirm that leaks from the transfer system will always be confined to fittings and flanges. A realistic failure scenario for a hose therefore needs to be considered.
All the testing evidence suggests that hoses will fail before rupture (with rupture pressures of 5-10 times the operating pressure) allowing the transfer to be stopped and the hose to be removed from service and disposed of.
The only possibility of a hose rupture is from a vessel or road tanker pulling away with the hose still attached. The IGF Code requirement for a break-away device means that a hose rupture is now a double jeopardy event: firstly, over-tensioning, and secondly, failure of the break-away valve. Double jeopardy events are not normally considered in safety assessments unless there is a common failure mode. In this hypothetical scenario, the failure mode is over-tensioning. The break-away valve is designed to operate in these circumstances so there is no common mode failure. In the light of this requirement, a hose rupture is not considered applicable.
SGMF’s emphasis has therefore been placed on determining a representative hole size for a hose failure. Within a Quantitative Risk Assessment (QRA), a range of release sizes would be analysed and appropriate frequencies assigned to each case. For the purposes of defining the Safety Zones within this study, a deterministic approach is taken which requires the selection of a single representative release size for a given transfer operation. This approach is also consistent with the “Deterministic assessment of the Safety Zone” described in ISO 20519.
Several options are available, as shown in Table D3 and Figure D2
© Society for Gas as a Marine Fuel 100
Figure D2: Hole size selection in hoses
0
5
10
15
20
25
30
0 50 100 150 200 250
hole
siz
e
Hose diameter mm
NL RIVM UK HSE Small
UK HSE large ISO
"TNO composite experiments" manhandling limit
© Society for Gas as a Marine Fuel 101
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Source Requirement
Dutch RIVM legislation Hole size is 10% of hose size.
Comment
RIVM uses a gradation of hole size with hose diameter. This suggests that worse events should happen with larger systems. The basis of RIVM is rubber hoses and therefore the 10% basis currently used in the report should be rejected as failure scenarios are completely different and guillotine failures have occurred.
The RIVM approach of a percentage failure up to some form of cap (50 mm in RIVM) on the linearity of the hole to hose diameter relationship may still be useful. Hoses at and above 6 inches are unlikely to be handled as they are too heavy, so bend radius limiters and other fatigue-limiting devices could be specified, resulting in a much reduced frequency of failure.
Source Requirement
DSM-2500003-RP-01 (basis unknown) 3 mm, 30 mm and full-bore rupture.
Comment
No further information on this Dutch code has been identified.
Source Requirement
UK Health & Safety Executive (HSE) Fixed hole sizes of 3 mm and 15 mm and full-bore rupture.
Comment
The UK HSE looks at two failure scenarios for hose failure, fixed holes sizes of 3 mm and 15 mm. A catastrophic/guillotine failure case is also given and alternatively described as a 25 mm hole. The basis of this work is unclear but is probably based on engineering judgements of metal hoses (small, about 25 mm) in chlorine service. The work is primari-ly on failure rates with no explanation given of the hole size choice. Credit is given for additional safety measures in the reductions of frequency.
Table D3: Hose failure hole size selection methods
© Society for Gas as a Marine Fuel 102
Table D3 continued: Hose failure hole size selection methods
Source Requirement
Statistical Review and Gap Analysis of LNG failure Rate Table (DTPH56-15-T-00008)
Comment
US onshore LNG regulators take the view that pipework of 6 inches and above cannot catastrophically fail and the maximum hole size should be 2 inches. Below this diameter complete failure of the largest connecting line is taken as the hole size – for example, with a 4 inch line tying into a 6 inch line, the hole size would be 4 inch; if the line was 3 inch it would be a 3 inch hole.
Source Requirement
ISO 20519 25 mm hole size based on broken instrument connection.
Comment
ISO (and Shell) have taken a similar position to the US authorities in suggesting that a failure of the largest connection to the hose should be considered, in this case a 1 inch (25 mm) instrument line. Relief valve connections (particularly TRVs) would probably also need to be considered. The actual size would be defined by the piping specification on board the vessel. An upper limit on the pipe size to be considered may be needed as some tappings may approach hose diameter.
Source Requirement
Composite hose manufacturer Multiple 1 mm holes.
Comment
Claimed, but no information provided to substantiate.
© Society for Gas as a Marine Fuel 103
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Source Requirement
TNO hose experiments (Dutch research organisation)
2 inch composite hose after fatigue: 0.6 l/min at about 10 bar through 9 cm (inner) and 5 cm (outer) radial crack/split (possibly 0.0002 kg/s).
4 inch composite hose after fatigue: un-known l/min at about 6 bar through 11 cm (inner) and 5 cm (outer) radial crack/split.
2 inch metal hose (no external helix) after crush test 0.4-1.5 ml/min gaseous nitrogen at 8 bar (about 0.0001 kg/s).
Comment
Composite hoseFrom the very limited TNO experimental work, both 2 and 4 inch hoses failed with the same length tear in the outer fabric. Inner fabric tears were different. Strength of the film would seem to be the failure point. One interpretation would be that composite hose failure is independent of hose size but a consequence of the manufacturing process; this is supported by manufacturer comments.
No details were given of the width of the tear. Taking a variety of tear widths (1-5 mm or 2-10% of length) and converting the area to a circular hole would give a hole size of between 8 mm and 17 mm.
Composite hoses appear to fail at or near their flange connections. The insertion of the flange into the multiple layer hose wrap is a discontinuity and appears to be a weakness. On this basis composite hoses could be considered as flange leaks.
Metal hosesNo indication was given in the TNO report of the hole size for the metal hoses.
Source Requirement
Professor Strohmeier Hole size is a function of hose size.
Comment
German academic Strohmeier produced a correlation for hole sizes in hoses as a func-tion of hose diameter. The correlation, similar to RIVM, suggested a hole size relationship of 8-9% of diameter.
No evidence has been found of the research work; it is not known what type of hose the correlation refers to or how robust and relevant the research is to LNG.
Table D3 continued: Hose failure hole size selection methods
© Society for Gas as a Marine Fuel 104
Reliability assessment work on piping in the offshore oil and gas industry – for example, the Process Leak for Offshore installation Frequency Assessment Model (PLOFAM) project – and for failure in rubber LPG/oil hoses – for example the Dutch RIVM methodology – links hole size to pipe/hose size. SGMF has taken the same approach for this guidance on metal and composite hoses.
The only failure size advice available for metal hoses is from the UK Health & Safety Executive and is for metal hoses transferring chlorine. Its expert judgement was that hole sizes of 3 mm and 15 mm diameter were appropriate for calculating safety distances. Applying these holes sizes across the transfer hose sizes expected within the LNG bunkering industry would suggest that a hole size of 6% of the hose diameter would be appropriate. Industry experience suggests that this is very conservative.
There is less information available about the failure of composite hoses. Manufacturers suggest a series of 1 mm holes or a short, few millimetres long, tear. If the perimeter of a tear is considered to be the circumference of a hole, similar hole sizes to metal hoses can be suggested. The conservatism of the model is therefore maintained.
The hole sizes for hoses based on the 6% diameter ratio are summarised in table D4.
Table D4: Hole sizes in metal and composite hose transfer systems
Hose diameter Hole size, mm
2 inch / 50 mm 3
3 inch / 75 mm 4.5
4 inch / 100 mm 6
6 inch / 150 mm 9
8 inch / 200 mm 12
10 inch / 250 mm 15
© Society for Gas as a Marine Fuel 105
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
If the design of the hose can be demonstrated to give a significant reduction in the likelihood of releases from the hose itself then releases from the flanges and valves at either end of the hose will dominate the release distance. The cryogenic hose industry is marketing double-wall hoses, in which a failure of the inner hose is captured within the wall of the outer hose. If this is proved in service, then the hose leak scenario would be much reduced or even eliminated. Leak scenarios, however, remain but are now limited to leaks from flanges, valves and fittings. In this case the Safety Zone would be based on releases from these locations.
Data from the Process Leak for Offshore installation Frequency Assessment Model (PLOFAM) project has been used to generate a series of hole sizes for fixed pipework, valves, and so on based on pipe diameter. PLOFAM calculates the risk of a leak of a certain size occurring in various plant components, based on correlating offshore failure rate data as power laws. The PLOFAM data is given per year. It is assumed that failures reveal themselves when transfer is taking place, that is, no adjustment is made for the fraction of year the equipment is in use. These estimates are therefore cautious.
Two risk scenarios were evaluated a 1 in 1000 (1x10-3) and a 1 in 10,000 (1x10-4) chance of a hole of a specific size occurring in the defined bunkering system. If the risk goes down to say 10-4 then the allowable hole size would increase as the frequency of occurrence of the hole declines with increasing diameter. (The risk here is risk of failure, not the risk to an individual worker. The risks to a worker located close by will be significantly less, their chance of being harmed is probably of the order of 1% given a release).
The transfer system is made up of various components in varying numbers so a worst-case transfer system also had to be defined to allow the risks to be aggregated into a single “hole size”. An alternate, simpler transfer system was also examined as a sensitivity case. Hole sizes were increased to the next largest whole millimetre for conservatism. Table D5 shows the results of the analysis.
© Society for Gas as a Marine Fuel 106
Table D5: Equivalent hole sizes from components of fixed transfer systems
Hose diameter Complex worst-case transfer system
Alternative simpler transfer system
15 m pipe, 8 valves, 11 flanges with gaskets, 2 instruments
10 m pipe, 6 valves, 7 flanges with gaskets, 1 instrument
Hole size for 1x 10-3 risk, mm
Hole size selected, mm
Hole size for 1x 10-3 risk, mm
Hole size selected, mm
2 inch / 50 mm 2.79 3 2.00 3
3 inch / 75 mm 3.07 4 2.21 3
4 inch / 100 mm 3.33 4 2.39 3
6 inch / 150 mm 3.80 4 2.72 3
8 inch / 200 mm 4.25 5 3.02 4
10 inch / 250 mm 4.66 5 3.30 4
© Society for Gas as a Marine Fuel 107
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
Appendix E: BASiL Validation
Reviews of accuracy against the original phenomenological models (CORCE, JINXTR/JINX2P, HAGAR and LSMS) based on the experimental data have shown that the vast majority of the BASiL calculations result in safety distances within ±10% of the actual values. About 10 points are in the range 0-10% and 2 marginally exceed +10%. LNG composition appears to be the source of most of the exceptions where a single parameter, Net Calorific Value (also known as Lower Heating Value), is being used to cover several different effects.
BASiL was validated by running a series of test cases for a range of input values. The safety distances from the spreadsheet (BASiL) were subsequently checked against individual runs of the outflow and dispersion models for the same inputs. Most results were generated for the most likely range of operating conditions, but some more extreme cases were generated to assess the performance for larger safety distances.
The results are shown in Figure E1. The solid line is perfect agreement and the dotted lines are ±10% and allow for the fact that BASiL always rounds up the result to the nearest integer – for example, 9.1 becomes 10.
© Society for Gas as a Marine Fuel 108
Figure E1: Validation of BASiL against detailed models/experimental data
© Society for Gas as a Marine Fuel 109
train
ing
& c
ompe
tenc
een
viro
nmen
tal
tech
nica
lsa
fety
cont
ract
ual
A: Horizontal free jets B: Vertical free jetsC: Radius of downward leaks onto water D: Height of downward leaks onto waterE: Radius of downward leaks onto land F: Height of downward leaks onto land