fabig technical note 8

FIRE AND BLAST INFORMATION GROUP

TECHNICAL NOTE AND WORKED EXAMPLES

Protection of Piping Systems subject to

Fires and Explosions

Technical Note 8

• This document is a deliverable of the Fire and Blast Information Group (FABIG) • FABIG would like to encourage comment and feedback from its membership. If you have any

comments on this Technical Note or any other FABIG activities please address them to the FABIG Project Manager at The Steel Construction Institute

The information in this document is published with the intent of making it available to members of the Fire and Blast Information Group (FABIG). The information is available for use subject to copyright. The information presented here is expected to contribute to the further improvement in safety. However, FABIG, the SCI and the reviewers assume no responsibility for any errors in or misrepresentations of such information or any loss or damage arising from or related to its use. No part of this publication may be reproduced without the written permission of FABIG and the SCI.

The Steel Construction Institute, Silwood Park, Ascot, Berkshire, SL5 7QN, United Kingdom Tel: +44 (0) 1344 623345, Fax: +44 (0) 1344 622944

Protection of Piping Systems Subject to Fires and Explosions

ii FABIG Technical Note 8


FABIG Technical Note 8 iii

FOREWORD

This Technical Note has been prepared as one of the FABIG deliverables to FABIG members.

The work was prepared to fill gaps in existing knowledge on the protection of piping systems against hydrocarbon fires and explosions.

This document was written and compiled by Fadi Hamdan of The Steel Construction Institute with significant input from Bob Brewerton of Natabelle Technology, Sava Medonos of Petrellus and Paul Rattenbury of KBR.


iv FABIG Technical Note 8


FABIG Technical Note 8 v

EXECUTIVE SUMMARY

The principles of structural design to resist fire and explosions on offshore installations have been extensively described in the Interim Guidance Notes and subsequent FABIG Technical Notes. Much of this guidance is also applicable to land-based petrochemical installations.

Techniques for the design of offshore piping to resist explosions have been described in OTO 1999-046 Explosion Loading on Topsides Equipment, Part 1 Treatment of explosion Loads, Response Analysis and Design. FABIG Technical Note 8 seeks to combine this guidance with the design for fire and to extend the techniques to cover onshore plant design. Guidance on minimising hazard, assessing hazard, on pipe support and pipe rack design is also given. The work follows a review of current industry practice in this area.

It is neither practical nor necessary to design all pipes to withstand explosion and/or fire effects and a two-level criticality rating system is advocated for those pipes that do require treatment.

Three design and analysis methods for explosions are described. The basic Category 1 analysis is a simple static load and code check approach to ASME B31.3. Category 2 is an enhancement of the method, with SDOF analysis to determine and incorporate suitable Dynamic Load Factors. Category 3 is a full MDOF – NLFEA approach for more complex situations. Categories 1 and 2 are illustrated by worked examples.

Industry experience has shown that there are practical limits to the general level of explosion resistance that can be achieved with process piping. It is consequently necessary to take measures in the overall plant design and local layout to reduce probable explosion pressures. An extensive part of the guide is devoted to this aspect, the objective being to reduce hazard and consequence by implementation of the principles of inherently safe design. For onshore plants a major goal is to prevent the spread of explosion effects from one plot/unit to adjacent plot/units and to people beyond the plant boundary (Land-use planning aspects).

An important feature that distinguishes design of piping systems from design of structures is that structures respond principally to the overpressure effects of explosions whereas pipe systems respond mostly to drag (explosion wind). Guidance is therefore given on how to relate drag loads to design overpressure loads, especially where the latter are produced probabilistically. Inevitably, some reliance on CFD modelling is required and guidance on this area is also included.

The ability of pressure-relieving and depressurisation systems to safeguard pressurised systems is critically dependent upon the assumptions made about the type and size of the threatening fire and the consequential levels of heat flux that each process segment is likely to be subject to. The onshore and offshore industry has traditionally used the American Petroleum Institute’s Recommended Practices (API RP 520 and 521) when designing pressure relief systems to enable pressure vessels and associated pipework to withstand the effects of fire. The recommendations in API RP 521 are typical for conditions of low heat flux fires in a refinery or chemical plant. However, it is now widely recognised that, should process plant fitted with protected systems designed to API RP 521 or a similar standard be exposed to severe fires, such systems may be insufficient to prevent failure of the pressure system before the inventory has been safely removed. The next revision of API RP 521 will incorporate restrictions as to its applicability to low heat flux and non-impinging / non-engulfing fires.

The main principles of design of pressure systems to resist fires have been relatively recently described in the UK Institute of Petroleum Guidelines for Design and Protection of Pressure Systems to Withstand Severe Fires and in similar guidance by Statoil / Norsk Hydro / Scandpower in Norway. This subject was also addressed in the UK HSE Offshore Technology Report OTO 2000 051.

The above UK and Norwegian guidelines address the limitations of API RP 521 and provide an up-to-date guidance for the design of pressure systems to withstand severe fires.


vi FABIG Technical Note 8

The objective of this Technical Note is to provide a methodology on the design and the protection piping systems and piping supports on both offshore installations and onshore plants for fires and explosions. The guidance covers the methods used to carry out both simplified design checks and advanced non-linear analysis.

The original research, on which this document is based, was sponsored by the Health and Safety Executive and was carried out by The Steel Construction Institute. FABIG sponsored the production of the Technical Note.


FABIG Technical Note 8 vii

CONTENTS

Page FOREWORD iii EXECUTIVE SUMMARY v ABBREVIATIONS ix GLOSSARY xi SCOPE OF THIS DOCUMENT xx PART A OVERVIEW OF DISCIPLINE ACTIVITIES 1 1. INTRODUCTION 3 2. OVERVIEW OF DISCIPLINE ACTIVITIES 9

2.1 Introduction 9 2.2 Description of design process 9 2.3 Overview of Disciplines 11 2.4 Development of information with progress of project 15

PART B HAZARD ASSESSMENT AND PLANT LAYOUT 20 3. HAZARD ASSESSMENT 22

3.1 Principles of hazard assessment 22 3.2 Goal setting approach 28 3.3 Determination of criticality levels for piping 39

4. PLANT LAYOUT 42 4.1 Overall layout aspects 42 4.2 Offshore installations 42 4.3 Onshore plants 46 4.4 Projectile risk 51 4.5 Local layout aspects 51 4.6 Non-Conductive Materials and Minimising Electrostatic Sparking Risk 54

PART C DESIGN 56 5. EXPLOSION LOADS 58

5.1 Parameters affecting explosion loading 58 5.2 Methods used for determining the explosion loading 61

6. DESIGN OF PIPING AGAINST EXPLOSIONS 76 6.1 Introduction 76 6.2 Effects of explosions on pipework 76 6.3 Design flow charts 76 6.4 Acceptance criteria 96

7. FIRE LOADS 98 7.1 Introduction 98 7.2 Fire characteristics to be considered 100 7.3 Determination of heat fluxes 101 7.4 Interaction of Fires and Explosions 104

8. DESIGN OF PIPING AGAINST FIRES 106 8.1 Introduction 106 8.2 Effects of fire on pipework 106 8.3 Design of piping systems 111 8.4 Firewater and other essential safety systems 123 8.5 Pipe supports 123 8.6 Flanges, bolts and welds 123 8.7 Fire protection 124 8.8 Optimisation of fire protection 129


viii FABIG Technical Note 8

9. DESIGN OF PIPE SUPPORTS AGAINST FIRES AND EXPLOSIONS 132 9.1 Introduction 132 9.2 Types of piping supports 132 9.3 Guidance for ductile construction 136

10. TYPES OF PIPING AND MATERIAL PROPERTIES 138 10.1 Standards used for piping and piping material 138 10.2 Typical materials used 139 10.3 Required material properties for carrying out blast assessment 139 10.4 Required material properties for carrying out fire assessment 139

11. OUTSTANDING ISSUES 142 11.1 Explosions 142 11.2 Fires 142 11.3 Supports 143 11.4 Piping Materials 143

REFERENCES 144 APPENDIX 1 Response of piping system subjected to an explosion loading 157 APPENDIX 2 Response of piping system subjected to fire 193


FABIG Technical Note 8 ix

ABBREVIATIONS

AFC Approved for construction AFFF Aqueous Film Forming Foam AFP Active Fire Protection ALARP As Low as Reasonably

Practicable API American Petroleum Institute ASME American Society of Mechanical

Engineers B/D Blowdown BDV Blowdown Valve BLEVE Boiling Liquid Expanding Vapour

Explosion BS British Standard CFD Computational Fluid Dynamics COMAH (regulations)

Control of Major Accident Hazards (Regulations)

CHRM Cost of a hazard reduction measure

DAF Dynamic Amplification Factor DAL Dimensioning Accidental Loads DLB Ductility Level Blast DLF Inbound Dynamic Load Factor DLM Direct Load Measurement EDP Emergency Depressurisation EER Escape, Evacuation and Rescue EN Euronorm ESD Emergency Shutdown ESDV Emergency Shutdown Valve F&G Fire and Gas (detection) FABIG Fire And Blast Information GroupFAR Fatal Accident Rate FEA Finite Element Analysis FEED Front End Engineering Design FEM Finite Element Method FES Fire and Explosion Strategy FPSO Floating Production Storage and

Offloading Unit GOR Gas to Oil Ratio GRP Glass Reinforced Plastics HSE (discipline)

Health, Safety & Environment

HSE (body) Health & Safety Executive (United Kingdom)

IGN Interim Guidance Notes

IR Individual Risk IRPA Individual Risk Per Annum IP Institute of Petroleum (United

Kingdom) JIP Joint Industry Project JFRT Jet Fire Resistance Test LNG Liquefied Natural Gas LPG Liquefied Petroleum Gas LYS Lower Yield Strength NDT Non-Destructive Testing NLFEA Non-Linear Finite Element

Analysis NORSOK (standards)

The Norwegian offshore sector standards

NPD Norwegian Petroleum DirectorateP&ID Piping and Instrumentation

Diagram PAU Pre Assembled Unit PDF Pressure Distribution Factor PFEER Prevention of Fire and Explosion,

and Emergency Response Regulations

PFP Passive Fire Protection PSD Process Shutdown PSV Process Safety Valve QRA Quantitative Risk Assessment RAC Risk Acceptance Criteria RDLF Rebound Dynamic Load Factor RP Recommended Practice SCF Stress Concentration Factor SDOF Single Degree of Freedom SIF Stress Intensity Factor (am., ref.

Caesar II) SLB Strength Level Blast SMS Safety Management System TEMPSC Totally Enclosed Motor-Propelled

Survival Craft TN Technical Note TR Temporary Refuge UF Utilisation Factor UKOOA United Kingdom Offshore

Operators Association UTS Ultimate Tensile Strength UYS Upper Yield Strength


x FABIG Technical Note 8

VHRM Value of a Hazard Reduction Measure

am. American eng. English


FABIG Technical Note 8 xi

GLOSSARY

GENERAL

Emergency depressurisation (EDP)

Controlled disposal of pressurised fluids to a flare or vent system when required to avoid or minimise a hazardous situation [26].

Emergency shutdown (ESD)

Control action undertaken to shut down equipment or processes in response to a hazardous situation [26].

Integrity Wholeness, soundness. Operator Individual, partnership, firm

or corporation having control or management of operations of the leased area or a portion thereof [26].

(Platform) Module A self-contained process and structural unit forming part of platform topside [2].

(Refinery) Unit or process Unit in a chemical plant

A process system within a geographical area performing a specific processing function.

Redundancy The performance of the same function by a number of independent means [14].

Source of release Point from which flammable gas, liquid or a combination of both can be released into the atmosphere [26].

Validation (of software)

Confirmation that software performs as intended, but without examining the truthfulness of its results.

Verification (of method)

Confirmation that a method gives valid results.

HAZARD AND RISK As Low As Reasonably Practicable (ALARP)

ALARP expresses that the risk level is reduced (through a documented and systematic process) so far that no further cost effective measure is identified. (The requirement to establish a cost effective solution implies that risk reduction is implemented until the cost of further risk reduction is grossly disproportionate to the risk reducing effect.) [99].

Availability The proportion of the total time that a component, equipment or system is performing in the desired manner [14].

Consequence The meaning is either: The outcome of major accidents expressed as physical phenomena such as gas concentrations, thermal radiation levels and explosion overpressures.

Or: The number of injured people, fatalities, environmental damage, asset damage, loss of revenue, impact on publicity.

Control (of hazards)

Limiting the extent and / or duration of a hazardous event to prevent escalation [26].

Dependability The ability of an item that another item can depend on under specified conditions.

Domino effect Events participating in causal sequence.

Endurance The period of time during which an item of equipment maintains its functionality.

Escape Act of personnel moving away from a hazardous event to a place where its effects are reduced or removed [26].

Escape, evacuation and rescue (EER)

General term used to describe the range of possible actions including escape, muster, refuge, evacuation, escape to the sea and rescue / recovery [26].


xii FABIG Technical Note 8

Evacuation The planned method of leaving the installation in an emergency [26].

Failure The termination of ability of an item to perform a required function (BS 4778).

Fatal Accident Rate (FAR)

The Fatal Accident Rate (FAR) is the potential number of fatalities in a group of people exposed for a specific exposure time to the activity in question. Generally, the FAR is expressed as the probability of fatality per 100 million exposure hours for a given activity. The 100 million exposure hours represents the number of hours at work in 1000 working lifetimes.

Fire and explosion strategy (FES)

Results of the process that uses information from the fire and explosion evaluation to determine the measures required to manage these hazardous events and the role of theses measures [26].

Flammable atmosphere

Mixture of flammable gas or vapour in air, which will burn when ignited [26].

Frequency The number of occurrences per unit of time [14].

Functionality The ability of an item to perform its required function.

Functional requirements

Minimum criteria which must be satisfied to meet the stated health, safety and environmental objectives [26].

Hazard Potential for human injury, damage to the environment, damage to property, or combination of these [26].

Hazardous area Three-dimensional space in which a flammable atmosphere may be expected to be present at such frequencies as to require special precautions for the control of potential ignition sources [26].

Hazardous area classification

Division of installation into hazardous areas and non-hazardous areas and the sub-division of hazardous areas into zones [26].

Hazardous event Incident which occurs when a hazard is realised [26].

Ignition sources Any source with sufficient energy to initiate combustion [26].

Incipient cause Initiating cause (of a failure).Individual risk The frequency at which an

individual may be expected to sustain a given level of harm from the realisation of specified hazards [14]. The Individual Risk (IR) expresses the probability per year of fatality for one individual. It is also termed as Individual Risk Per Annum (IRPA). The IR depends on the location of the individual at a given time and his / her contents of work.

Inherent safety A system safe by its own in-built properties, characteristics or features.

Initiating failure frequency

A statistically obtained frequency of the starting failure.

Isolatable process segment

All piping and equipment within one depressurisation volume. The ESD or PSD valves connected to the segment define the limit of the depressurisation volume. A single pressure vessel, storage, or transportation tank, etc., can also be a process segment. During depressurisation the pressure in the process segment is reduced through the BDV and its depressurisation orifice [10].


FABIG Technical Note 8 xiii

Main safety functions

Safety functions that need to be intact in order to ensure that pollution is controlled and personnel that are not directly and immediately exposed, may reach a place of safety in an organised manner, either on the installation or through controlled evacuation. (Examples of main safety functions are main support structure, escape ways, control centre, shelter area (Temporary Refuge) and evacuation means.) [99].

Mitigation (of hazardous event)

Reduction of the effects of a hazardous event [26].

Performance standard

A statement which can be expressed in qualitative or quantitative terms, of the performance required of a system, item of equipment, persons or procedure, and which is used as the basis for managing the hazard, e.g. planning, measuring, control or audit – through the life cycle of the installation [16].

Pressure system Pressure vessels and their associated pipework, valves, flanges and other equipment [10].

Prevention (of hazardous event)

Reduction of the likelihood of hazardous event [26].

Probability A number in a scale from 0 to 1 which expresses the likelihood of occurrence of an event [14].

Quantitative risk analysis

The quantified calculation of probabilities and risks without taking any judgement about their relevance [14].

Quantitative risk assessment

The quantitative evaluation of the likelihood of undesired events and the likelihood of harm or damage being caused together with the value judgements made concerning the significance of the results [14].

Reliability The ability of an item to perform a required function under stated condition for a stated period of time (BS 4778).

Rescue The meaning is either: Process by which those who have entered the sea directly or in survival craft are retrieved to a place where medical assistance is available [26].

Or: The process of the transfer of injured personnel to a place of relative safety.

Residual Risk Remaining risks after risk has been reduced to ALARP.

Risk Combination of the probability that a specified undesired event will occur and the severity of the consequences of that event [26].

Safety barrier (essential safety system)

Any system which has a major role in the control and mitigation of fires and explosions and in any subsequent EER [26].

Safety goals The objectives of effort to protect personnel from hazards.

Safety Management System (SMS)

The management arrangements that are concerned with safety performance. (A SMS includes a safety plan to implement the safety goals.)

Societal risk The relationship between frequency and the number of people suffering from a specified level of harm in a given population from the realisation of specified hazards [14].

Survivability The ability of an item to survive an event, but with a reduced functionality after the event.


xiv FABIG Technical Note 8

Temporary Refuge Place provided where personnel can take refuge for a predetermined period whilst investigations, emergency response and evacuation preplanning are undertaken [26].

Totally enclosed motor-propelled survival craft (TEMPSC)

Craft capable of sustaining the lives of persons in distress from the time of abandoning the installation [26].

Zone (area classification)

Volume around the source of release to the point where the flammable atmosphere has been diluted by air to a sufficiently low level [26].

EXPLOSION

Blast wave A pressure pulse formed by an explosion [14].

Boiling Liquid Expanding Vapour Explosion (BLEVE)

The sudden rupture due to fire impingement of a vessel / system containing liquefied flammable gas under pressure. The pressure burst and the flashing of the liquid to vapour creates a blast wave and potential missile damage, and immediate ignition of the expanding fuel-air mixture leads to intense combustion creating a fireball [14].

Blockage The fraction of an area which does not provide open venting of free passage for a flame [2].

Confined explosion An explosion of a fuel-oxidant mixture inside a closed system (e.g. vessel or module) [14].

Congestion The high density of equipment and pipework in a plant.

Deflagration The chemical reaction of a substance in which the reaction front advances into the unreacted substance at less than sonic velocity. Where a blast wave is produced which has the potential to cause damage, the term explosive deflagration is used [14].

Detonation An explosion caused by the extremely rapid chemical reaction of a substance in which combustion is caused by the pressure rise and the reaction front advances into the unreacted substance at greater than sonic velocity [14].

Direct Load Measurement (DLM) (method)

An analysis in which the differential pressure on an object is found by subtracting the downstream pressure from the upstream pressure, rather than by considering the drag loading on the obstacle. The method also covers situations where several upstream and several downstream monitor points are used.

Ductility Level Blast (DLB)

The design for extreme (10000 year return probability or higher) explosion event. Also known as the dimensioning explosion loading.

Explosion Violent combustion of a flammable gas or mist, which generates pressure effects due to confinement of the combustion-induced flow and / or the acceleration of the flame front by obstacles in the flame path [26].

Physical explosion Explosion arising from the sudden release of stored energy such as from failure of a pressure vessel, or high voltage electrical discharge to earth [26].


FABIG Technical Note 8 xv

Explosion epicentre

The location in the gas cloud where explosion pressure is highest: it is scenario dependent.

Explosion scenario A specific gas cloud size, stoichoimetry, location and ignition position.

Far field (shock wave)

The far field is a location outside the module or plot/unit where the explosion actually occurs. If the distance from edge of the exploding gas cloud is sufficient, impulses will have a vertical shock front (zero rise time).

Fire / explosion area

An installation is divided into explosion or fire areas. Usually an area is bounded by fire/ blastwalls or open spaces (fire breaks). But in extensive process plants and FPSO’s the area might be the limit of the extent of the envelope of gas clouds/fires that would occur as a result of a leak fire/ occurrence within the area.

Monitor point In a CFD analysis, a local point selected for monitoring explosion characteristics (pressure, impulse, drag etc).

Near field (shock wave)

The impulse that occurs within or close to the exploding gas cloud.

Panel pressure In CFD analysis, explosion pressures can be averaged over a given panel size. Peak pressure and impulse are then the peak value of the average pressure over the panel and the impulse is the time history of this pressure.

Strength Level Blast (SLB)

The design for strength level blast. The dimensioning load to assess the static strength of the structure i.e. load for which response is elastic.

Vapour Cloud Explosion

The preferred term for an explosion of a cloud made up of a mixture of a flammable vapour or gas with air in open or semi-confined conditions [14].

Windage area The shadow area exposed to incident wind and is “A” in the expression for drag force Fd = 0.5 Cd A ρ V2.

FIRE

Active Fire Protection

Equipment, systems and methods which, following initiation, may be used to control, mitigate and extinguish fires [26].

Area (firewater) deluge

Firewater deluge where firewater is evenly distributed over a floor area.

Blowout (ignited)

A form of jet fire resulting from ignition of a high pressure flow of oil and / or gas issuing from an uncontrolled well, possibly with a substantial fallout of heavy fractions which may also ignite on or around the installation as pool fires [2].

Dedicated (firewater) deluge (also equipment protection)

Firewater deluge where firewater is directed to specific items of equipment.

(Firewater) deluge system

System to apply firewater through an array if open spray nozzles by operation of a valve on the inlet to the system [26].

Deluge system (water deluge system)

System to apply firewater through an array of open spray nozzles by operation of a valve on the inlet to the system [10].

Depressurisation (blowdown)

Controlled reduction of pressure by disposal of fluids inventories from a process segment (normally to the flare or vent system), sometimes referred to as blowdown [10].


xvi FABIG Technical Note 8

Depressurisation - automatic

Depressurisation that is initiated directly and automatically from the fire and gas (F&G) detection system. The signal may be via the ESD system. In some cases time delays are introduced in order to allow isolation valves to close prior to start of the depressurisation operation [10].

Depressurisation - manual

Depressurisation initiated by an installation operator. If depressurisation is initiated automatically by a high level ESD and that level of ESD requires operator action, this is still considered manual depressurisation. Manual depressurisation can be initiated either by a master push button on a matrix, by individual push buttons on a matrix or via the computer consoles [10].

Depressurisation system

System designed to enable depressurisation (blowdown). Normally this consists of piping connected to the process segment, depressurisation valve (BDV) with associated actuator, instruments, etc., and orifices, tail pipe, flare / vent headers, knock-out drum and flare stack and tip [10].

Film boiling (heat transfer)

Boiling of liquid contents of a pipe or a vessel limited to the inner surface of the pipe / vessel wall, which creates a vapour film between the liquid and the wall surface.

Fire area A hazardous area where accidental fire may occur, limited by boundaries such as firewalls beyond which fire should not escalate.

Fire intensity The heat flux emanated by the fire flame or flame temperature.

Fire endurance The time during which the object affected by fire remains functional.

Fire scenario The information describing the flammable substance being released, fire type (jet, pool fire, etc., ventilation or fuel-controlled, obstructed or unobstructed), fire intensity, flame length, fire duration, etc.

Fire test A test designed to quantify the resistance to fire of elements of construction and / or materials applied to the element. Conventionally, such tests are carried out in purpose-built furnaces operating to a defined time / temperature curve [2].

Firewall A designated partition which, by nature of its construction and certification status, is warranted to resist a standard or hydrocarbon fire test for a particular time period. The term can apply to a floor or roof panel [2].

Flash fire The combustion of a flammable vapour and air mixture in which flame passes through that mixture and negligible damaging overpressure is generated [14].

Fuel-controlled fire

A fire in which the rate of fuel consumption is controlled by the rate of supply of fuel to the fire, rather that the availability of oxygen for combustion [2].

Jet fire Ignited release of pressurised, flammable fluids [26].

Nucleate boiling (heat transfer)

In the nucleate boiling regime, bubbles are formed on the heated surface, they are detached from the surface and dissipated in the liquid.


FABIG Technical Note 8 xvii

Passive Fire Protection (PFP)

Coating or cladding arrangement or free-standing system which, in the event of fire, will provide thermal protection to restrict the rate at which heat is transmitted to the object or area being protected. [26].

PFP coatback Fire protection coating extending from a main pipe to branch pipe or from a structural primary member to a secondary member that prevents the temperature in the main pipe / primary member to rise to unacceptable level from the heat conducted through the branch pipe / secondary member.

Pool fire Combustion of flammable or combustible liquid spilled and retained on a surface. [26].

Running liquid fire

Fire involving a flammable liquid flowing over a surface. [26].

Stiffness The force / displacement relationship of an element; the ratio of stress to strain. [2].

Two-phase flow A flow in which both the liquid and vapour or gas phases co-exist within close proximity [2].

Ventilation-controlled fire

A fire in which the combustion rate is controlled by the availability of oxygen rather than the supply of fuel [2].

View factors The proportion of the field of view of a receiving surface that is filled by a flame [2].

PIPEWORK LOADING AND RESPONSE Explosion loading and response

Ductility Ratio The maximum deflection of a pipe’s span divided by the deflection of the same reference point at the elastic limit.

Fracture, brittle (brittle fracture)

A break initiating from a crack without ductile deformation capacity prior to a complete break.

Fracture, ductile (ductile fracture)

A break initiating from a crack with ductile deformation capacity prior to a complete break.

Inbound response The dynamic response of the pipe in the direction of the applied explosion loading.

Limit, deformation (deformation limit)

Capacity for deformation or plastic straining ( expressed as a distance or a percentage strain).

Limit, ductility (ductility limit)

Limit of ductile deformation beyond which fracture, rupture or rapid loss of resistance due to buckling occurs.

Limit, strain (strain limit)

A limit set on allowable strain to ensure that fracture, rupture, or buckling collapse does not occur. The strain limit may incorporate safety factors to account for uncertainties or material variability.

Load factor, dynamic (Dynamic Load Factor (DLF))

1) For elastic quasistatic design, the factor by which the peak load applied to a member has to be multiplied in order to give the same peak stresses and deflections as the peak dynamic response of the member to the applied load impulse.

2) For elasto-plastic response it is the design quasistatic strength divided by the peak applied load, whose value is such that the peak deflection of the member would be the same as would be found in a full time domain dynamic response due to the applied load impulse.


xviii FABIG Technical Note 8

Load factor, dynamic, rebound (Rebound Dynamic Load Factor (RDLF))

The factor by which the peak load applied to a member has to be multiplied in order to give the same peak stresses and deflections due to rebound deflection back towards the applied loading.

Pressure, drag (drag pressure)

The wind pressure that occurs due to the flow of gas, air or burnt explosion products past a reference obstacle or member = 0.5 ρ V2.

Pressure, field (field pressure)

The barometric pressure measured at a reference point.

Pressure, headline (headline design pressure)

The basic peak explosion pressure used for designing the module or plot / unit, usually the 10-4 return value, or DLB value. Also known as the dimensioning peak pressure.

Rebound Response The dynamic response of the pipe in the direction of the applied explosion loading.

Stress Intensity Factor (am., Caesar II)

Stress Concentration Factor (eng.).

Fire loading and response

Brittle fracture Fracture of the material of construction without any reserve due to ductility [2].

Conduction (thermal conduction)

The mode of heat transfer in which energy exchange takes place from the region of high temperature to that of low temperature by the kinetic motion or direct impact of molecules, as in the case of fluid at rest, and by the drift of electrons, as in case of solids.

Convection (heat transfer)

Heat transfer associated with fluid movement around a heated body; free convection: warmer, less dense fluid rises and is replaced by cooler, more dense fluid; forced convection: the fluid movement is forced by external force such from a fan or pump [2].

Creep The time dependent straining of material. It is measured under a constantly applied load and, in the context of this report, at a fixed temperature [2].

Emissivity A constant used to quantify the radiation emission characteristics of a flame. Emissivity of a perfect blackbody is 1 [2].

Material yield stress

The stress at which a steel sample departs from linear elastic behaviour to plastic deformation in a standard tensile test [2].

Plasticity The behaviour of steel following yield [2].

Plastic deformation The deformation which occurs following yield [2].

Specific heat (capacity)

The amount of heat energy required to heat-up 1kg mass of solid or fluid by 1K, cp (J/kg K).

Strain hardening (work hardening)

The tendency of an elastic-plastic material to exhibit increased resistance at high strains [2].

Thermal absorptivity

A measure of an overall emissivity in the radiative heat transfer from the flame to the surface of a body, comprising the flame emissivity and the emissivity of the body surface. Thermal absorptivity is a product of flame emissivity and body surface emissivity.

Thermal capacity The ability of a material to store heat [2].


FABIG Technical Note 8 xix

Thermal conductivity

The rate of heat flow by conduction in a given direction is proportional to the area normal to the direction of heat flow and to the gradient of temperature in that direction. The proportionality constant is called the thermal conductivity of the material, k (W/m K).

Thermal diffusivity A measure for heat propagation into the medium during changes of temperature with time. Thermal diffusivity is a ratio of thermal conductivity and the product of density and thermal capacity.

Upset conditions Excursion from plant operating conditions into temperatures and pressures that are higher/lower than those corresponding to normal operating conditions.

METHODS

Computational Fluid Dynamics (CFD) (method)

A mathematical model in which the region of the flow is subdivided by a grid into a large number of control volumes [2].

Section method (also called Hp/A method)

A method of temperature calculation based on the relationship between the heated surface area and heated volume.


xx FABIG Technical Note 8

SCOPE OF THIS DOCUMENT

This document gives guidance on the protection of process piping and piping supports against hydrocarbon fires and explosions on offshore installations and land-based petrochemical plants. Included also is guidance for active fire protection piping.

“Process piping” in the context of this guidance covers pipes, pipe supports, flanges, valves including valve actuators, and fittings used for processing and handling chemical fluids. Vessels, vessel nozzles and supports are not included.

It should be recognised from the outset that this guidance provides an ‘opinion’ on the state-of-the-art practice that has been formed based on liaising with different industry experts and based on ‘perceived’ experience from several previous projects. It should be noted, however, the guidance provided in this document has not been formally validated for the protection and design of piping systems against fires and explosions. The document is therefore by definition interim and is expected to be updated once more industry information and / or feedback is received.

This document is divided into the following three main parts:

• Part A provides an overview of discipline activities;

• Part B provides guidance on Hazard Assessment and Plant Layout Issues; and

• Part C provides guidance for design against fires and explosions.

This document uses information from regulations, standards, codes of practice, guidelines, conference papers and meetings of professional bodies. The main sources of information are listed below.

Blast and Fire Loading and Response

• Phase I Blast and Fire Engineering Project for Topside Structures [1];

• Interim Guidance Notes [2]; • Phase II Report [3];

• CMR Explosion Handbook [4]; • Design of Offshore Facilities to Resist Gas

and Explosion Hazard [5]; • Explosion Loading on Topsides Equipment,

Part 1 OTO 1999 046 [6]; • FABIG Technical Note 6 High Strain rate

and elevated temperature data [7]; • Review of the response of process vessel

and equipment to fire attack [8]; • Guideline for the Protection of Pressurised

Systems Exposed to Fire [9]; • Institute of Petroleum Guidelines on

protection of pressurised systems against Fires [10];

• ASME B 31.3 Code, Process Piping [11]; and

• Explosion Pressure Evaluation in Early project Phase [12 and 13].

Hazard Management Philosophy and ALARP Guidance

• UKOOA Guidelines for Fire & Explosion Hazard Management – 1995 [14];

• HSG 65 – Successful Health & Safety Management – 1997 [15];

• HSE L65 Prevention of Fire & Explosion & Emergency Response – 1997 [16];

• A Guide to Offshore Installation Regulations [17];

• HSE R2P2 - Reducing Risks , Protecting People – 2001 [18];

• HSE Guidance on ALARP Decision for Offshore Division Inspectors [19];

• HSE Hazard Management in Structural Integrity

OTO 98 148 [20]; OTO 98 149 [21]; OTO 98 150 [22]; OTO 98 151 [23]; and

• Guidance on ‘As Low As Reasonably Practicable’ in COMAH [24].

These documents also identified gaps in the knowledge and understanding. This Technical


FABIG Technical Note 8 xxi

Note is, therefore, by definition, interim and will become outdated and require updating as these gaps continue to be closed by the industry.

This document considers almost exclusively explosions and fires based on hydrocarbons as a fuel, originating from process equipment.

This Technical Note is written for disciplines involved in safety engineering. It also should be of interest to all of following disciplines:

• Safety engineering; • Pipe stress engineering; • Piping supports engineering / draughtsmen; • Layout engineering; • Structural engineering; and • Process engineering.

The reading of the whole document will provide the reader with a multi-discipline overview. The Table below shows recommended reading for each discipline.

Extent Discipline

recommended reading

Further reading

Safety engineering All Sections Piping stress engineering

Section 2 Section 3.4 Section 3.6 Section 8 Section 9

Section 3 Section 4 Section 7

Piping supports engineering / draughtsmen

Section 2 Section 6.3 Section 9

Section 6 Section 8

Layout engineering Section 2 Section 3 Section 4

Section 3 Section 5 Section 6

Structural engineering

Section 2 Section 5 Section 6 Section 7 Section 8 Section 9 Section 10

Section 3 Section 4

Process engineering

Section 2 Section 3 Section 4 Section 6 Section 8

Section 5 Section 7

Pipework with hazardous contents or with a function to fulfil in emergency, which is potentially exposed to explosion or fire, forms one of safety essential systems. The design of safety essential systems is a multi-discipline task, where the systems should respond in harmony with the response by personnel and vice versa. Safety engineering has the central function to ensure that this happens and should be aware of methods that are available to achieve it. It is therefore recommended that safety engineering disciplines read all Sections of this Technical Note.


FABIG Technical Note 8 1

PART A OVERVIEW OF DISCIPLINE ACTIVITIES


2 FABIG Technical Note 8



1. INTRODUCTION

This document addresses the design methodology and organization of the design process for the design of piping systems and their supports for resistance to explosion and fire.

The document has been prepared on the basis of most recent regulations, standards, codes of practice and guidance, and also design practice in the petrochemical industry both onshore and offshore. Extensive references have been made to the sources of information.

Wherever the information is related to a specific type of installation and plant it is noted in the document. It is the authors’ experience, however, that a substantial number of methods and practices used in the onshore processing industry may also be used in the offshore industry and vice versa.

Explosion Pipes respond to explosion wind (drag pressure) rather than directly applied field pressure. Resistance to explosion wind demands that pipes be strong, that supports be located at close spacing along the pipe and that the structures on which they are mounted have adequate strength.

Direct shock loading on pipes is small compared to explosion wind. This is mainly due to the speed with which shock fronts can pass piping systems, allowing rapid equalisation of upstream and downstream pressures.

Pressure and shock loads can however cause significant movement (and occasionally collapse) of deck structures, walls and large equipment items. If pipes interconnect these items, such that the pipes themselves become strained between adjacent supports, then it becomes essential for them to have the necessary ductile deformation capacity. However, design for ductile deformation capacity can conflict with the requirement for strength to resist explosion wind.

Another conflicting factor is that operating requirements are best met by having flexibility in the piping systems and maintainability requires that piping equipment can be removed for maintenance.

A further conflicting factor is that the use of increased supports leads to additional structures whose sole purpose is to support pipes. This is turn may lead to increased congestion, and associated increases in the probable explosion pressures.

Experience from offshore projects has shown that there is a practical limit to the explosion wind and fire resistance that can generally be provided and a blanket design quasi-static explosion wind pressure is normally established and applied solely to what are deemed to be the most critical pipes. The value has historically been in the range 0.15 to 0.3bar and is still considered to be a realistic limit.

Fire The response of pipes containing chemical fluids to fire is complex. The heat is transferred from the fire by radiation and convection onto the pipe surface, fire protection coating or thermal insulation. The heat is then conducted through the wall of the pipe and is transferred to the pipe contents. Pressure in the pipe is normally reduced by a pressure relief or a depressurisation system. This process is, however, counter-acted by the increase of the pressure due to the heating-up of the pipe contents, boiling and thermal expansion of the fluids inside the pipe. The strength of construction material reduces with rising temperature and the applied stress may exceed the material strength, resulting in pipe rupture. All these processes vary with time.

For pipes containing chemical fluids, all the above processes have to be simulated using multi-physics approach. Simplistic approach based on the relationship between the surface area receiving the heat and the mass being heated-up may be used for pipe supports. The simplistic approach may also be used for flanges and valves, where the effects of fluid contents are negligible.

The obvious solution is the use of fire protection coatings. There is a conflict, however, between providing passive protection of pipes and piping equipment yet maintaining access for inspection and maintenance and avoidance of latent (corrosion) defects.

Again, flexibility is an important issue so that pipes can move due to operating thermal expansion or in fire, and this may be in conflict



with the desire for strength, and for additional supports, to react against explosion wind.

Criticality Certain lines are very critical and are required to remain intact during the explosion and/or fire event, in order to control and limit the potential escalation of the event. A typical offshore installation or land-based petrochemical plant may have more than 1000 potentially hazardous lines and it is not practical to apply a rigorous treatment to all of them.

In practice not all pipes need to be assessed, i.e. the failure of some categories of lines during an explosion or fire event would not contribute significantly to the overall severity of the event or to the ability to recover from the event. In all cases material loss will be an inevitable consequence of a major explosion, either during the explosion or the subsequent fire, but good attention to detail and overall layout can keep the loss within acceptable limits.

Current practice employs a “Criticality Rating” approach to determine which lines should be provided with explosion or fire resistance and the associated level of design loading that should be applied. This is covered in Section 3.2 of this Note. Another common practice is to employ a sectionalisation philosophy, dividing the installation into areas of varying loadings / hazardous content. These concepts are in line with the requirements of the main goal for the installation / plant of reducing risks from hazards to As Low As Reasonably Practicable (ALARP) through inherent safety principles (see Sections 3.2.1 and 3.2.8).

Design codes and analysis - explosions The primary loading element to which pipes respond is drag (explosion wind). However, the plant and structure design, Safety Cases both offshore and onshore, guidance on tolerance limits for process equipment, use the term of field pressure. Different methods for determining the drag loading are discussed and related to the level of risk in the installation. These range from simplified methods for translating field pressure into equivalent drag pressure and zoning the platform into various levels of drag pressure. For installations with high risk, more advanced methods, e.g. Computational Fluid Dynamics (CFD) methods, are used. This aspect is covered in Sections 5.2.4

and 5.2.5. Once the loading is determined, it has to be incorporated within the piping design process.

The basic design code for the design of piping systems is ASME B31.3. In this code, explosion loading can be treated as an “occasional load” with overstress allowed. Piping stress analysis is usually computerised and a commonly used tool is Caesar II. ASME B31.3 is a working stress code but it has essential features for allowing for some forms of non-linearities such as stress intensification factors for tees and bends. Furthermore, allowable safe limits on externally applied loads for combination with line pressure and temperature are stipulated. Section 6 describes the application of this method and Example 1 in the Appendix shows how it can be applied to a typical critical piping system.

The ASME B31.3 / Caesar II approach is fundamentally a quasi-static method, whilst explosion loading is dynamic and causes a dynamic response. Section 6 shows how equivalent static loads can be developed from dynamic load information so that the quasi-static method can be applied. In most cases the design quasi-static load has to be larger than the peak dynamic load due to dynamic amplification of response. It should be noted that, historically, dynamic analysis has rarely been performed and Dynamic Amplification Factors (DAF) are effectively assumed as unity. This shortcoming is to some extent covered by adopting the conservative working stress approach in ASME B31.3. A discussion on the conservatism or otherwise of this approach is provided in Section 6 (Sections 6.3.1 to 6.3.3 in particular).

In individual cases it is possible to design for higher loadings than those corresponding to the capacity of the piping as determined by quasi-static methods. In such cases, it will usually be necessary to determine design piping loads by a dynamic method and this sometimes involves considering the compliance of the structures on which the pipes are mounted (e.g. pipe racks on FPSOs). It is also possible to enhance the dynamic resistance of piping systems by employing a variety of measures that will increase strength and ductility. These aspects are covered in Sections 6.3.5 to 6.3.8.

Design codes and analysis - fires General guidance is provided in the IP Guidelines and the Statoil - Norsk Hydro -



Scandpower guidance, which are both complemented by OTO 2000 051. All these documents describe the multi-physics processes that are to be taken into account in the determination of the internal pressure in the pipework affected by fire.

Hydrocarbon, petrochemical and chemical processing plants, as with pressure systems, are either designed to withstand the highest expected pressure given by the codes or fitted with means of preventing over pressurisation. The protection of pressurised systems is usually a detection and shut-down system, pressure relief valves (PRV) and / or bursting discs, and emergency depressurisation (EDP) (blowdown). These are designed to limit the maximum pressure within the systems and to prevent catastrophic failure, or to reduce the risk and consequences of failure.

The onshore and offshore industry has traditionally used the American Petroleum Institute’s Recommended Practices (API 520 and 521) when designing pressure relief systems to enable pressure systems to withstand the effects of fire. However, it is now widely recognised that they were not originally intended to cover all foreseeable fire scenarios. Should process plant fitted with protective systems designed to API RP 521, or a similar standard, be exposed to a fire, such systems may be insufficient to prevent failure of the pressure system before the inventory has been safety removed. The API RP 521 is currently being revised, its limitations will be stated in the next revision and a reference will be made to the IP Guidelines.

The IP and Scandpower guidance requires the application of a multi-physics method and calculation of stress in the pipe.

The multi-physics method is a time-dependent method of calculating the time history of temperature and pressure in the system taking account of the following:

• The heat received by the pipe, conducted through the pipe wall and transferred to the pipe contents;

• The temperature rise of the pipe; • The thermodynamics of the pipe contents

including the phase change; • Masses of the fluid compositions inside the

pipe;

• Pressure reduction due to pressure relief or depressurisation counter-acted by the pressure rise from the heating-up of the pipe contents, evaporation, boiling and thermal expansion of the vapours and liquid; and

• The reduction of material strength due to the rising temperature.

A simplistic multi-physics analysis represents a pressure system as a straight pipe and also calculates applied stress and time to rupture. The stress-raising effects of branches, bends, supports, etc. may be included in the form of stress concentration factors, but, the effects of thermal expansion of the piping system, pipe supports, bellows and other pipe-system related effects are not included.

In order to include the pipe-system effects the pressure and temperature results at selected time instances from a multi-physics analysis may be transferred to a linear elastic pipestress analysis, such Caesar II. Caesar II will compute static stress at the time instant of the pressure and temperature loads. As the heating-up process is slow relative to structural effects, structural dynamic effects are ignored.

Alternatively, the instantaneous pressure and temperature from the multi-physics analysis may be transferred to a non-linear pipestress analysis and the elasto-plastic response beyond the first yield may be investigated.

The multi-physics analysis also calculates the temperature drop in the depressurisation (blowdown) pipe due to the rapidly expanding stream of gas through blowdown valve. This is performed to check possible material embrittlement or for the selection of appropriate material.

The multi-physics method uses stress as a criterion. A Section Method (often called as “Hp/A” method) is normally used for the calculation of temperature rise of structural members and temperature criteria, but it cannot be used for piping systems with internal fluids, unless it can be proved (by a multi-physics analysis) that the pipe internal pressure does not exceed the design pressure, where the latter must be related to the elevated temperature throughout the duration of the fire. The Section Method can be used for pipe supports.



The cooling effects of fluids inside the pipe are misleading and should be ignored. The pressure variation is governed by the thermal expansion of the pipe contents, evaporation, boiling and the pressure increase from these effects.

All these aspects are addressed in Section 8.

Layout issues Explosions with more than 1.5bar field pressure will have extensive zones with higher drag pressures than 0.15 to 0.3bar. In such cases piping equipment may be extensively damaged, which may be further compounded by a follow-on fire. It is desirable to avoid locating sensitive pipes (such as gas compressor piping) in such areas. Another goal is to employ inherently safe design principles (through good global and local layout methods, to avoid such high pressures arising). This raises a large number of local and overall layout issues, which if optimally applied can:

a) reduce the likelihood of explosion and fire events;

b) reduce the severity of the events; and c) reduce the likelihood of escalation of the

event and overall consequence.

These layout issues have to address both explosion and fire as it is the combination of the two effects that will dictate the overall hazard frequency and consequence severity. Section 4 describes some potential measures, both for offshore platforms and onshore plants; all based on experience gained on projects during recent years.

It should be recognised that in some cases, and for some scenarios, it may not be feasible to employ good local / global layout policies (such as avoiding gas compressor piping in high drag zones). In such instances a “barrier philosophy” needs to be applied to limit the spread of damage in an individual event and avoid domino effects in large installations (escalation between modules and plots). The two basic barrier methods are containment by walls or distancing by open spaces. The objectives are to limit the risk of domino effects and limit the risk of an explosion opening up large inventories of hydrocarbons and, of course affecting inhabited locations. These are discussed in Section 3.2.8 and in Section 4.

Pipe supports need to be strong enough and compliant enough to support the pipes connected to them and permit the necessary movements of the pipes. Section 6.3.4 provides guidelines on pipe support design, traditionally an area of neglect. It is recommended to have a criticality rating system for pipe supports and to have criticality 1 (top level) supports designed by persons reporting to the structural group, rather than the lower-level activity normally associated with standard supports. Examples of typical pipe supports are given in this Technical Note.

Design Methodology - explosions The design methodology linking most of the relevant design aspects is shown in the design flow chart in Figures 6.1 and 6.2. It is broken down into a set of basic activities, described in Section 6.2, consisting of:

• Layout optimisation; • A criticality rating system to identify where

design effort should be focused; • Methods for the determination of quasi-

static design loads; • The basic ASME B31.3 – Caesar II quasi-

static design procedure; • Criteria for establishing when pipes need to

be subject to more enhanced dynamic analysis;

• Guidance on how dynamic analysis can be performed (2 levels, SDOF and MDOF); and

• Pipe supports and pipe rack design.

Design methodology – fire The design methodology for pipework exposed to fires is outlined in Figures 8.4 and 8.5. The design process can be broken down to the following activities:

• The preparation of fire protection philosophy;

• Process and safety design including sectionalising and pressure relief / depressurisation systems;

• Hazard evaluation; • The determination of the pipework

criticality level; • The determination of detailed fire loading; • Pipework response analysis; and



• The consideration of mitigation options if the design is not adequate, and re-design.

Application examples The design process is illustrated by two examples:

1) A critical hydrocarbon pipe in an offshore separator area subject to explosion, using ASME B31.3 and Caesar II; and

2) The same pipe subject to fire, using multi-physics analysis and the combination of multi-physics analysis and Caesar II.

These examples are presented in Appendices 1 and 2 respectively.



2. OVERVIEW OF DISCIPLINE ACTIVITIES

2.1 Introduction This Section provides an overview of the various disciplines that are involved in the protection of piping systems against fires and explosions. Section 2.2 provides a discussion of the design process, while Section 2.3 gives a brief discussion of the responsibilities of each discipline involved in the design of piping against fires and explosions. Section 2.4 discusses the availability of information at various stages in the project.

2.2 Description of design process

2.2.1 Basic design goals The basic goal of the piping design process is:

• To make critical hydrocarbon piping systems to survive credible explosion and fire events;

• To prevent escalation of the initial explosion and / or fire event to neighbouring modules or units, and

• To make fire fighting piping systems to survive credible explosion events and remain functional in the fire events.

2.2.2 Explosions Pipes are affected primarily by explosion wind (drag) loads but deck (support) movement due to

blast or fire is sometimes an important consideration.

Large obstacles and obstacle groups dam the flow of burnt products and receive high transient loading due to differential pressure.

In the far-field, shock waves act on pipes but their direct effect is usually small. The shock loads however do displace the larger items to which pipes are connected and this is also to be a design consideration for the piping, especially where differential movement between interconnected components or systems is concerned.

In an onshore process plant explosion or on an FPSO the pressure wave accelerates through the expanding gas cloud, increasing in intensity as it propagates, see Figure 2.1 below.

Factors governing explosion pressure are:

• The size of the combustible portion of gas cloud;

• Equipment and pipework congestion; • Blockage of the flow of combustion

products; and • The spacing and width of run-down gaps

(open fire breaks) between equipment zones.

Figure 2.1 Acceleration of pressure wave through an expanding gas cloud

zone of weaker back flow

Zone of forward drag flow

Direction of travel of explosion wave

Gas cloud

Blast wall



2.2.3 Fires Pipes are affected by heat from fires which will heat-up fluids inside the pipes and, due to thermodynamic effects, this may cause the internal pipe pressure to rise despite the activation of pressure relief valve and / or emergency depressurisation.

The heat intensity of the fire may reduce by the activation of firewater deluge. The success of such an action will depend on the reliability of the firewater system.

Flanges that have no fire protective coatings will lose their tightness and new leaks will develop; where the inventory is flammable, it may be ignited by the fire.

The rate of pressure relief / depressurisation may be increased by increasing the orifices, however, there may be limits in the flaring capacity.

The obvious solution is the application of fire protection coatings, together with a moderate increase of pressure relief / depressurisation orifices. The use of PFP coatings has, however, a number of disadvantages, such as the increase of corrosion risk, requirement for inspection and maintenance of the coatings, and lengthy removal and re-application of the coatings for pipework maintenance and inspection.

Firewalls and also fully plated decks will prevent flame penetration to neighbouring areas.

Empty non-safety-essential pipes with internal atmospheric pressure pose negligible risk when affected by impinging flame. The pipe would rapidly heat up and the air inside the pipe expand, which would cause the pressure relief valve to open and reduce the pressure. In the unlikely event that the pipe ruptures, there will be no release of flammable fluids and therefore no potential for escalation.

2.2.4 Main factors affecting piping design practice

The following measures should be adopted whenever possible to reduce the hazard frequency and consequence.

Reduction of hazard frequency The number of potential release points should be minimised by using fully welded pipework in preference to flanged connections. It should be

noted that this solution may not be the most optimal for process modifications required in the later life of an offshore installation, which may be required due to changes of the reservoir composition. Welding on the platform normally requires the stoppage of production and hence loss of revenue. An alternative may be to use a habitat for welding operations on a live platform, but past experience with welding habitats has been very costly in some cases. Also, fully welded pipework requires more time for installation and maintenance, and therewith, longer exposure of personnel to hazards.

Currently the best solution seems to be the use of welded pipework combined with compact flanges. The main characteristic of a compact flange is the flange face geometry. It includes a slightly convex bevel with the highest point, called the heel, adjacent to the bore and a small outer wedge around the diameter of the flange face.

During make-up of the connection, the bevel is closed and flange face-to-face contact is achieved. Most of the bolt pre-tension is transferred as compressive forces between the flange faces at the heel, while a minor compressive force is transferred through the outer wedge. The back face of the flange is parallel to the flange face in order to prevent bending of the bolts in the assembled condition.

The seal ring for the connector is a flexible metal ring, which is located in a groove. The groove is located close to the outer diameter of the connecting pipe. In the made-up condition the outer wedge acts as an external barrier and keeps the flange faces and bolts out of contact with the external medium.

Only a minor part of the total pre-tension is transferred through the “elastic” flexible seal ring. The flange face-to-face contact after make-up combined with the small bevel of the flange faces assures that the flange faces are static during operation, i.e. no relative flange face-to-face displacement occurs at design loads. The connector behaves as a rigid joint (monolithic) at design load levels. The use of compact flanges offers the following advantages:

• The initiating leak frequency is lower than that of ordinary API flanges. A correctly made-up flange will have a leakage probability in the same order of magnitude as a welded connection. This results in a



reduced probability of accidental releases, explosions and fires;

• The strength of compact flanges is equivalent to threaded connections. In riser systems, for example, compact flanges can be designed to be at least as strong as the adjoining pipe;

• Compact flanges are lighter than API flanges. This results in reduced loading of supporting structures and smaller member cross-sections, leading to smaller aerodynamic diameter and reduced explosion overpressures;

• Compact flanges are smaller than ordinary API flanges, which results in reduced congestion and blockage, and smaller aerodynamic diameter, which leads to reduced explosion overpressures; and

• The diameter of a compact flange is smaller than that on an ordinary flange. This leads to narrower pipe racks, less congestion and blockage, which results to lower explosion overpressure.

A compact flange has fatigue properties better than the adjoining weld and there is very little stress variation in the bolts even for very high external cyclic load. On the other hand, compact flanges require accurate installation, which in turn requires accurate locations and dimensions of the points where compact flanges will be attached. The measurements of these locations and dimensions may be carried out using techniques such as laser and photogrammetry.

For modifications on offshore installations the combination of welded pipework with compact flanges may be combined with pre-fabrication of parts of process segments onshore with a final installation offshore. This reduces the exposure duration of personnel to platform hazards.

The probability of explosion and / or fire may be reduced by improved ventilation, as this reduces the probability of the formation of flammable cloud or flammable mixture, and ignition. This can be influenced by ventilation-friendly sub-division of the installation into fire/ explosion areas. Fire areas should be the same as explosion areas, with realistically designed fire-rated blast walls that will minimise the likelihood of escalation to neighbouring fire/ explosion areas.

The probability of fatalities and injuries in case of a explosion and/or fire hazard materialising

should be reduced by a harmonised response of personnel and essential safety systems defined in a realistic emergency response plan.

Reduction of consequence severity Explosions The following measures should be adopted wherever possible to reduce the loading on piping systems, and to ensure that the piping will respond in a ductile manner to explosion loading:

• Reduce headline design pressures: this is particularly important since reducing the design overpressure leads to proportionately larger reductions in the drag wind loading acting on the piping (see Section 2.3.4);

• Shelter pipes behind beams, which will avoid them being subjected to high wind loading (see Sections 4.2 to 4.4); and

• Provide ductility to piping systems, which will ensure that the piping can deform and dissipate energy under dynamic loading, without failure to the piping supports or the flanges (see Sections 6.3.5 to 6.3.9).

Fires The following measures should be considered to reduce the fire loading and response severity of piping systems:

• The use of water deluge to reduce heat flux from the fire should be considered, taking into the account the reliability of the deluge system; and

• Fire protection coatings should be optimised to achieve minimum coatings and minimum pressure relief / blowdown orifice.

2.3 Overview of disciplines One of the main design goals in the design of piping and piping supports against fire and explosion hazards is the prevention of escalation. Failure of a small pipe or vessel can lead to an escalation of an initial explosion event and the outbreak of fires, which in turn can lead to the failure of larger vessels and / or pipes, and to more severe consequences, as shown in Figure 2.2 [25]. Barriers should be provided in the design to prevent such escalations. Safety management systems provide barriers against such escalations. The proper design of piping systems to withstand explosions and fires is one



such barrier against escalation. Another example of an escalation barrier (or “essential safety system”) is the use of fire / blast walls. An organisational barrier is a Safety Management System.

The protection of piping systems to withstand fires and explosions requires the involvement and interaction of several engineering disciplines. The responsibilities of, and interaction between, these disciplines is discussed in following Sections.

Figure 2.2 Effect of piping failure on

potential for escalation [25]

2.3.1 Process engineering discipline The responsibilities of the process engineering discipline include:

1. Carrying out the initial design of pressure relief / depressurisation systems;

2. Design of sectionalisation of the process system;

3. Carrying out the changes to the design of sectionalisation and pressure relief / depressurisation systems; and

4. Liaison with other disciplines.

It should be noted that the pipe sizes are determined by the process engineering discipline based on process simulation of operating conditions.

2.3.2 Safety engineering discipline The responsibilities of the safety discipline include:

1. The preparation of hazard management plan and liaising with all disciplines to ensure that the plan is properly implemented;

2. The preparation of fire and explosion protection strategy (sometimes also termed as “fire and explosion protection philosophy”). This includes but is not limited to the definition of safety essential systems (barriers), piping and piping support criticality and prevention of escalation and the determination of fire scenarios;

3. Making sure that all relevant disciplines are involved in the design process as appropriate;

4. Carrying out risk assessments and associated studies, including the determination of explosion and fire loads;

5. Jointly with other disciplines, identifying remedial solutions for the resolution of unacceptable cases;

6. Input to installation layout and other design input alongside with other project disciplines;

7. Input to the process engineering on sectionalisation and pressure relief/ depressurisation design;

8. Input to the specification of fire protection coatings for pipework and supports;

9. Input on fire proofing to pipework and structural drawings;

10. Participation in regular design reviews;

11. Carrying out interdisciplinary cross checks to confirm the design integrity; and

12. Liaising with other disciplines to ensure that the response of pipework affected by explosion and / or fire is harmonised with the response of other safety essential systems and personnel.

2.3.3 Piping stress engineering discipline

The responsibilities of the piping stress engineering discipline include:

1. Preparing a computer model for the critical piping systems and liaising with the safety



discipline to incorporate the blast load as a quasi-static load within the computer model (e.g. Caesar II);

2. Incorporating the blast load into the relevant load cases, carrying out a linear analysis and screening the results to identify which piping systems should receive further consideration;

3. Incorporating the fire loads and carrying out the fire response analyses, assessment of the results and identifying remedial solutions for the resolution of unacceptable cases;

4. Liaising with the structural, piping support and layout disciplines to ensure that the supports and the substructure are designed for the loads from the piping system; and

5. Liaising with the piping layout discipline to ensure that local piping layout has been optimised to reduce fire and explosion induced stresses in critical piping systems.

2.3.4 Piping layout engineering discipline

The responsibilities of the piping layout engineering discipline include:

1. Liaising with the piping and structural disciplines to ensure that local piping layout is optimised to reduce the risk of piping failure, while allowing for ductility, thermal expansion and strength requirements;

2. Liaising with piping, structural and safety disciplines to ensure that global layout is optimised to reduce the overpressure and drag loading due to explosions; and

3. Liaising with piping, structural and safety disciplines to ensure that the global layout is optimised and is consistent with the compartmentalisation / barrier philosophy that is employed on the plant / installation to prevent the escalation of explosions and fires.

2.3.5 Piping support engineering / draughting discipline

The responsibilities of the piping support discipline include:

1. Liaising with the piping, structural and safety disciplines to ensure that the piping system can resist the differential displacement due to support movement;

2. Liaising with safety and structural discipline to ensure that the structure can withstand the blast loading passed through the support to the substructure; and

3. Designing safety critical supports to resist the associated loading criteria due to fire / blast loading.

2.3.6 Structural engineering discipline The responsibilities of the structural discipline include:

1. Liaising with piping stress engineers and carrying out simplified dynamic analysis to determine dynamic amplification ratios for piping systems as requested by piping stress engineers;

2. Liaising with piping stress engineers and carrying out multi degree of freedom analysis for critical piping systems, where necessary;

3. Designing the supporting structure carrying the piping systems and piping support to resist dynamic blast loading from these systems; and

4. The preparation of fire protection drawings for pipework supports.

Figure 2.3 illustrates the involvement and activities of each discipline in the design of pipework against fire and explosions at various stages of design. The deliverables at the end of the detailed design in this context are As-For-Construction (AFC) drawings and associated specifications. As there may be some minor modifications made during the construction phase, this information needs to be updated and included in the As-Built drawings and specifications.



Figure 2.3 Responsibilities and interaction between disciplines at various stages of the design life cycle

Process Safety Structural Piping

Layout Pipe

Stress Pipe

Supports

Process: Determine pipe sizes, equipment, vessels, pressure, temperature. Prepare PFDs.

Safety: Carry out Concept Safety Evaluation. Determine safety functions and barriers.

Structural: Review the locations of blast and fire walls.

Piping Layout: Prepare piping layout drawings.

All Disciplines: Liaise with each other.

Process: Further develop and refine data on piping, equipment, vessels, pressure and temperature. Further develop PFDs. Prepare

P&IDs. Safety: Prepare fire and explosion protection philosophy. Further develop and refine data on safety functions and barriers. Carry out

Risk Assessment. Determine explosion and fire loads. Prepare explosion and fire protection specifications. Structural:

Review explosion and fire loads. Determine the capacities of blast and firewalls. Piping Layout:

Prepare FEED piping layout drawings. Pipe Stress:

Perform piping rupture analyses. All Disciplines:

Liaise with each other.

Process: Further develop and refine data on piping, equipment, vessels, pressure and temperature. Further develop and refine PFDs and P&IDs.

Safety: Update fire and explosion protection philosophy. Further develop and refine data on safety functions and barriers. Refine Risk Assessment. Refine

explosion and fire loads. Update explosion and fire protection specifications. Structural:

Review explosion and fire loads. Refine the capacities of blast and firewalls. Piping Layout:

Prepare detailed piping layout drawings. Pipe Stress:

Review piping rupture analyses and carry out additional ones as required. Pipe Supports:

Design of pipe supports for explosions and fire. All Disc iplines:

Liaise with each other.

Conceptual

FEED

Detailed design



2.4 Development of information with progress of project

A typical offshore or onshore project will have a considerable number of pipes and piping supports, of varying levels of criticality. The last update on congestion and blockage is obtained from the final layout at the tail end of the design and the time available for design iterations against explosions may be rather short. Therefore, where possible, experience from previous projects should be used to ensure that the design based on intermediate data is acceptable.

2.4.1 Explosion data with progress of project

As can be seen from the above discussion, piping design against fires and explosions requires interaction and exchange of information between several disciplines. This interaction between disciplines is needed not only for the design of piping against fires and explosions, but also at a higher levels in the design process for building an accurate explosion model that reflects realistic congestion layouts.

Figure 2.4 shows the building and exchange of information between the piping, layout, safety and design disciplines, with the aim of defining congestion in the explosion model as a project is progressing [12 and 13]. Initially large equipment and pipes, together with the preliminary structural layout, is used to build the topside CAD model, which in turn is used when generating the first explosion pressure analysis. The explosion loading is used for the first structural analysis and the equipment and piping design. The CAD model is updated based on the new structural layout, equipment and piping design, and a new explosion pressure analysis is carried out. Results from the second explosion pressure analysis are used to carry out a second structural response analysis and a further analysis on pipes and equipment. Results from the latter two are fed back into the CAD model to build the final CAD model. It should be noted that the pipe support stiffnesses used in the analyses may have a significant effect on the pipe stresses and therefore should be checked against assumed values.



Figure 2.4 Interaction between piping and other disciplines to incorporate congestion into explosion model [12 and 13]

At the beginning of the project, very few details regarding the piping that will be used on the topsides are available. This is particularly true for the small and medium size piping that can significantly contribute to the degree of congestion, and hence the total overpressure developed due to an explosion event. Therefore, if the piping geometry is added to the CAD model only when the actual information becomes available, this may result in cases where the total final overpressure is higher than the overpressure that was used at the beginning of the project to design the structure and the major pipes and equipments. To avoid such situations, where the overpressure is grossly underestimated at the beginning of the project, it is useful to introduce artificial piping and equipment congestion models.

Nonetheless, at the beginning of the project a factor should be included to allow for lack of information that is necessary for the accurate estimate of the quantity of anticipated congestion.

The proper use of such an approach can lead to initial estimates of overpressure very close to the final overpressure value, and thus avoid the need for re-assessment or strengthening.

Congestion models would vary according to platform size, type and function. One such congestion model, based on OTO 1999 048 [12], is shown in Figure 2.5.



Block Volume of

Block Causing Congestion (m3)

Number of bocks

Total Volume (m3)

Function / Description of Block Causing Congestion

1 50 4 200 Main pipe rack in the process area

2 55 2 110 Transverse pipe rack 3 50 12 600 Manifold piping 4 20 11 220 Equipment piping (E-W) 5 17 4 68 Equipment piping (E-W), well

head area 6 20 5 100 Equipment piping (N-S) 7 104 6 624 Secondary under-deck piping

in the process area 8 80 4 320 Secondary under-deck piping

in the well head area 9 120 2 240 Secondary under-deck piping

in the manifold area 10 50 4 200 Main pipe rack in well head

area TOTAL 2682

Figure 2.5 Typical Congestion Model for an Offshore Module [12]

First the platform is divided into separate areas by function. Then for each area the need for main, secondary and tertiary pipe racks is identified, and estimates for pipe specification, sizes and length are provided. In this manner it is possible to build artificial congestion models that are used to determine an accurate estimation of overpressure during the early stages of the project.

OTO 1999-048 concentrates on the artificial congestion methodology for the piping elements but the method is normally extended to include valves, fittings, cable racks, secondary and tertiary structure.

2.4.2 Fire data with progress of project Initial process data in the form of fluids, pressures, temperatures and the volumes of isolatable inventories are normally available from conceptual or FEED phases of the project. Although the accuracy of this information is at the conceptual or FEED levels, the fire type and duration can be rapidly estimated together with preliminary fire loads based on tabulated values.

This information is used in the preparation of the initial fire protection philosophy and made up to date as the design progresses and the data becomes more accurate. Data updates are obtained from process engineering, which influence fire scenarios and protection. Changes in fire intensity and duration will affect fire protection requirements for pipes. For flanges,

wellhead area process areautility area

N3

1 0 1 0 1 1

3

2

2

4

4

4

4

5 5

6

6

6

7

77

88 9

9

5 5

6

6



however, it may be assumed that an initial fire will cause additional leaks from flame-impinged unprotected flanges as these lose their tightness within 5 minutes after the start of the fire; the vast majority of realistic fires will be of duration greater than 5 minutes.

The process is illustrated in Figure 2.6. At the completion of the design the designed pipework should respond in a manner reflecting realistic fire scenarios, and the pipework response should be in harmony with the response of other safety essential systems and personnel.

Figure 2.6 Flowchart illustrating the interaction between disciplines in the development of piping fire protection on a project

Process Safety Pipe stress

Piping layout

Process input: Fluids PFD’s P&IDs Pipe sizes Equipment Vessels Pressure Temperature

Fire protection philosophy, including: Fire hazards Fire scenarios Preliminary fire loads

Process design: Sectionalisation Pressure relief Depressurisation

Hazard evaluation

Determination of pipe criticality

Detailed fire loading

Analysis of fire

Fire protection requirements and specifications

Fire protection input to P&ID drawings

Fire protection input to piping layout drawings

AFC documentation



PART B HAZARD ASSESSMENT AND PLANT LAYOUT



3. HAZARD ASSESSMENT

This Section describes:

• Principles of hazard assessment; • Goal setting approach; and • Determination of critical levels for piping.

Section 3.1 addresses the hazard design philosophy and the relationship between hazard assessment, hazard identification and risk analysis. Section 3.2 describes the goal setting approach and its application throughout the plant lifecycle, and the corresponding safety management systems and performance measures. It describes how this philosophy is applied throughout the life cycle of the platform using inherently safe design principles. The procedures used for deciding when the risk becomes tolerable are also described in Section 3.2, together with control and mitigation measures for the residual risk. Section 3.3 describes the philosophy for the determination of critical levels of piping.

3.1 Principles of hazard assessment

3.1.1 Hazard assessment process A hazard is defined as the potential for human injury, damage to the environment, damage to the property, or combination of these [26]. It would be unproductive to assess all hazards in detail and analysis effort is therefore focused on major hazards identified by the process of systematic Hazard Identification (HAZID).

HAZID may require some calculations of the consequences of the hazard being realised, which is often called Hazard Analysis (HAZAN). Some risk practitioners also use the term HAZAN for the analysis of severity of hazardous event and its consequences on people and plant within QRA.

Hazard assessment comprises HAZID, cause identification and examination and risk assessment. As such it is an overall term for the identification of major hazardous events and the quantitative evaluation of the likelihood of these events, and the likelihood of harm or damage

being caused together with the value judgement made concerning the significance of the results.

Most process hazards are related to the consequences of accidental release of hazardous substances where the hazardous properties of the substances are not exacerbated by the hazard materialising. However, some pressure systems (mainly onshore) contain pressurised reactive chemicals. The involvement of these chemicals in a fire could initiate runaway reactions; these would need to be taken into consideration in the hazard assessment.

When relevant, an accident scenario is developed for each initiating event. The scenario defines the nature of the accident and the subsequent chain of consequences. Figure 3.1 gives an example of consequence chain.

The consequence chains can be conveniently represented in the form of an Event Tree. The use of Event Tree involves first the estimation of the frequency of the initial event or hazard that triggers the problem. Each branch of the Event Tree represents an additional consequence chain eventually resulting in a series of outcome events.

For a hazard that arises from the accidental release of hydrocarbons, the initiating leak frequency from each section of the process can be estimated using either a “parts count” approach with generic component failure frequencies or historical data on leak frequencies. In order to use historical data, it is necessary to be sure that there is no significant under-reporting and that leaks are reported in sufficient detail to be able to estimate release size.

An analysis of the consequences resulting from any event is an important aspect. The majority of effort involved in the consequence analysis is devoted to the examination of hydrocarbon releases with the estimation of outflow and release effects, where the latter predominantly involves estimation of the magnitude of fires and explosions and their effects on plant and structures. This may be achieved on the basis of physical models, an analysis of historical statistics, research or design, information on previous accidents, knowledge of the behaviour of pipework, equipment and structures affected by fires or explosions, and the experience of experts.



The frequency of the initiating event and probabilities of the follow-on events (consequences) constituting the branch points are put into the Event Tree and the frequencies of the outcome events are calculated. The follow-on event probabilities are often called as branch probabilities. They are evaluated on the basis of “rule sets”, where the latter are determined based on the expected consequences of the events on pipework, equipment and structures.

Tables 3.1, 3.2, and 3.3 gives data on the damage of plant components and structures exposed to fires or explosions that may be useful for the establishment of the rule sets.

It should be noted that the information in Tables 3.1 and 3.2 is only approximate, as it does not include data such as the explosion overpressure impulse, the surface area receiving the pressure load and connection types for structures. Work by Salzano and Cozzani [27] presents an overview of further data reported in the literature for damage to process equipment caused by explosion overpressure, identified discrepancies in this data and sources of the discrepancies.

The failure of pipework, vessels, equipment and structures exposed to fire depends on duration of the exposure. As shown in Table 3.3, it is possible to give estimated time to failure for some components, whilst pressure equipment and structures normally require calculations in order to determine the time to failure.

Failure modes of pipework and pipework components affected by explosions and fire are described in Sections 6.2 and 8.2 respectively.

Figure 3.2 illustrates the process of assessment of hazards and their evaluation against risk criteria. Risk is defined as a combination of the chance that a specified undesired event will occur and the severity of the consequences of that event [26]. As such, risk assessment in Figure 3.2 comprises frequency and consequence analyses, risk picture, and risk evaluation.

Hazards should be eliminated or prevented in every project in accordance with the relevant laws and regulations. The hazard elimination and prevention starts at concept selection. Once the preferred concept is chosen and justified the focus turns to reduction of residual risk, however small, that still exists in the concept after the elimination / prevention process. Residual risks are further reduced through FEED and detailed design to an As Low As Reasonably Practicable (ALARP) level.

The hazard assessment is often supplemented by Safety Integrity Level (SIL) analysis. The purpose of SIL analysis is to establish whether plant control systems are designed in proportion to the level of expected risks.



Table 3.1 Reported damage to piping and equipment from explosions [4]

Overpressure (barg)

Equipment 0.

03

0.07

0.10

0.14

0.17

0.20

0.24

0.27

0.30

0.34

0.37

0.41

0.44

0.48

0.51

0.54

0.60

0.61

0.65

0.68

0.82

0.95

1.09

1.2

1.36

>1.3

6

Cooling tower 2 4 3

Tank: cone roof 3 9 19

Fired heater 5 7 18

Chemical reactor 1 ? 14 18

Filter 6 4 20 18

Regenerator ? ? 14 15

Tank: floating roof 9 19 3

Pipe supports 14 17 13

Gas meters 15

Electrical motors 6 10 20

Blower 15 18

Fractionation column

16 18

Horizontal pressure vessel

14 7 18

Extraction column 7 20 18

Stream turbine 10 11 17 20

Heat exchanger 10 18

Tank sphere 8 7 18

Vertical pressure vessel

7 18

Pump 10 20 Notes: 1. Windows and gauges break, 2. Louvers fall at 0.3 to 0.5 psi; 3. Roof collapses; 4. Damage to inner parts; 5. Brick cracks; 6. Projectile damage; 7. Unit moves and pipe breaks; 8. Bracing fails; 9. Unit uplifts; 10. Power lines severed; 11. Controls damage; 12. Block wall fails; 13. Frame collapses; 14. Frame deforms; 15. Case damaged; 16. Frame cracks; 17. Piping breaks; 18. Unit overturn and/or collapse; 19. Unit uplifts; 20. Sliding/rocking motion


FABIG Technical Note 8

Table 3.2 Data reported in the literature for damage to process equipment caused by explosion overpressure [28]

Damage Overpressure (bar)

5% window shattering 0.005

50% window shattering 0.02

Collapse of roof of a storage tank 0.07

Connection failure of corrugated panelling 0.07 - 0.14

Minor damage to steel framework 0.08 - 0.1

Wall of concrete blocks shattered 0.15 - 0.2

Collapse of steel framework 0.2

Collapse of self-framing steel panel building 0.2 - 0.3

Ripping of empty oil tanks 0.2 - 0.3

Small deformation of pipe bridge 0.2 - 0.3

Big trees topple over 0.2 - 0.4

Panelling torn-off 0.3

Displacement of pipe bridge, failure of piping 0.35 - 0.4

Damage to distillation columns 0.35 - 0.8

Collapse of pipe bridge 0.4 - 0.55

Loaded train wagons overturned 0.5

Brick wall (0.2 to 0.3m thick) shattered 0.5

Movement of round tank, failure of connecting piping 0.5 - 1.0



Table 3.3 Time to failure of pipework, vessels, equipment and structures affected by fire [29]

Failure Fire Scenario Time to Failure Pipe rupture Flame impinging onto pipe with no fire

protection. Determine the time to failure by multi-physics analysis as described in this Technical Note.

Excessive deformation of pipe supports leading to loss of tightness and potential rupture

Flame of heat flux of 250kW/m2 impinging onto a pipe support with no fire protection.

<5min

Flange, loss of tightness Flame of heat flux of 250kW/m2 impinging onto a flange with no fire protection.

<5min

Hub connector or clamp flange, loss of tightness

Flame of heat flux of 250kW/m2 impinging onto a connector or flange with no fire protection.

<5min

Valve, loss of tightness Flame of heat flux of 250kW/m2 impinging onto a valve with no fire protection.

<10min

Safety valve, opens at a pressure lower than the setting pressure

Flame of heat flux of 250kW/m2 impinging onto a safety valve with no fire protection.

<10min

Bursting disc, opens at a pressure lower than the setting pressure or is destroyed

Flame of heat flux of 250kW/m2 impinging onto a bursting disc device with no fire protection.

<10min

Pressure vessel rupture with the potential formation of projectiles

Flame impinging onto pressure vessel with no fire protection.

<40min depending on the flame size with respect to vessel size, vessel contents, wall thickness and the size of pressure relief/blowdown orifice. Determine the time to failure by multi-physics analysis as described in this Technical Note.

Pressure vessel rupture with the potential formation of projectiles

Flame impinging onto a pipe attached to a pressure vessel. The pipe is unprotected and the vessel is protected so that heat is conducted by the pipe into the pressure vessel shell forming a hot spot with loss of strength.

<40min depending on the size of the pipe and fire intensity.

Excessive deformation of vessel supports leading to loss of tightness at nozzle flanges

Flame of heat flux of 250kW/m2 impinging onto a pipe support with no fire protection.

<5min

Loss of loadbearing capacity of a structural member, which may lead to large deformation in some locations and loss of tightness of pipework

Flame of heat flux of 250kW/m2 impinging locally onto a structural member with no fire protection.

<15min depending on the member size.

Collapse of structure or its part leading to loss of tightness of pipework and large releases of hazardous fluids

Flame of heat flux of 250kW/m2 impinging locally onto a joint of structural members or engulfing several joints.

<30min depending on the member sizes.

Collapse of atmospheric storage tanks, road tankers, rail tank cars and marine tankers leading to large releases of hazardous fluids

Flame impinging onto the storage or transport tanks with no fire protection.

<40min depending on the flame size with respect to tank size, tank contents, wall thickness and the size of the pressure relief device. Determine the time to failure by multi-physics analysis as described in this Technical Note.

Note: Time to failure for heat fluxes other than 250kW/m2 should be determined by transient calculations


FABIG Technical Note 8

Figure 3.1 Example of consequence chain following release of flammable gas

Hazard assessment

planning

System description

Hazard identification

Frequency analysis Consequence analysis

Risk calculation

Risk evaluation

Further risk reducing measures

Risk acceptance criteria

Risk reducing measures

PART OF SAFETY MANAGEMENT AND RISK CONTROL

HAZARD ASSESSMENT

RISK ASSESSMENT

Figure 3.2 Illustration of the hazard assessment process

3.1.2 Level and type of risk assessment As outlined in the HSE guide to the Offshore Installations (Safety Case) Regulations - 1992 [17] and the HSE guidance on ‘As Low as Reasonably Practicable’ (ALARP) Decisions in Control of Major Accident Hazards (COMAH) Regulations [24], the level and type of risk assessment should be determined using the proportionality concept.

Proportionality must be considered in at least two aspects of the risk assessment:

• The robustness of the risk assessment used; and

• The depth of the ALARP demonstration. Proportionality may also be appropriate when considering the concept of gross disproportion used in assessing the adequacy of the ALARP demonstration.

Pressure vessel rupture

Fire Ignition Explosion Release of flammable gas Ignition

Damage of pipework and further release of flammable gas



The HSE guidance for the COMAH regulations states that the depth of the analysis in the operators risk assessment should be proportionate to the scale and nature of the major accident hazards presented by the establishment and the installations and activities of it. The depth of analysis that needs to be present depends on the level of risk predicted before any additional risk reduction measures are applied. The nearer the risk to the intolerable boundary, the greater the depth of analysis needed. Various methods of risk assessment may be used depending on proportionality. These range from qualitative at the lowest level, through semi-quantitative up to quantitative at the highest level of risk.

In order to select the appropriate method of analysis a rapid assessment of the level of risk is first required. This should be based on qualitative factors such as presence of hazardous materials, level of manning on platform, etc, considered in HAZID.

In Table 3.4, levels of frequency and consequence are defined in a Risk Matrix according to factors such as level of manning, presence of hazardous material, etc. These are discussed in more detail in the UKOOA guidance [14] and ISO standards [30].

Table 3.4 Risk Matrix as a function of frequency and consequence

Frequency Consequence

Low Medium High

High Medium High High Medium Low Medium High Low Low Low Medium These considerations can be included in the HAZID process. The level and type of risk assessment can then be determined following the HAZID and prior to the start of the risk assessment process.

3.2 Goal setting approach It is useful to place the blast and fire strategy for piping within the broader context of the fire and blast hazard management plan for the platform, and in turn, to place that within the overall hazard managements system and to relate that to the goal setting approach used by the process industry. To this end, it is useful first to introduce the following definitions:

Safety goals The objectives of effort to protect personnel from hazards.

Safety Management System (SMS)

The management arrangements that are concerned with safety performance. (A SMS includes a health, safety and environment (HSE) plan to implement the safety goals, where the safety plan includes a fire protection philosophy.)

Performance standards

The standards of performance expected from essential safety systems.

Figure 3.3 shows the relationship between the above three definitions, which are discussed in more detail in Sections 3.2.1 to 3.2.4.

Figure 3.3 Goal setting approach

3.2.1 Defining safety goals The main design objective is to reduce the risk from hazards to as low as reasonably practicable (ALARP). Fires and explosions are two of the hazards for which this statement applies, and piping systems on platform topside facilities and process plants are one issue that must be considered when considering fire and explosion hazards. For the purpose of reducing the risk from fire and explosion hazards to as low as reasonably practical, the UKOOA guidelines



[14] and COMAH regulations [31] identifies the following aims:

• Identify, analyse and understand all fire and explosion hazards and associated effects;

• The risk corresponding to fire and explosion hazards identified above should be as low as reasonably practicable;

• A suitable order of priority, and a suitable combination, of prevention, detection, control and mitigation systems for fire and explosion hazards should be implemented and supported throughout the life cycle of the offshore platform;

• The above prevention, detection, control and mitigation systems should have performance measures proportionate to the required risk reduction;

• The design, operation and maintenance of the above prevention, detection, control and mitigation systems should be carried out by competent personnel; and

• Any changes that may occur throughout the lifecycle of the installation, and that may affect the likelihood and / or consequence of any fire or explosion hazard event (and therefore may make the risk on the installation deviate from an ALARP state) should be identified and assessed. The prevention, detection, control and mitigation systems should be modified and updated as necessary to take into account any such changes.

The following additional goals, relating to piping systems, can be inferred from those outlined above:

• Prevent events leading to high explosion overpressures, or

Minimise the failure frequencies and probabilities leading to severe explosions

Minimise of the consequences of severe explosions

• Prevent the occurrence of severe fires, or

Minimise the failure frequencies and probabilities leading to the occurrence of severe fires

Minimise the consequences following from the occurrence of severe fires

• Prevent the failure of safety critical piping in case of explosions and fires, or

Minimise the failure frequencies and probabilities leading to failure of critical piping

Minimise the consequences following from failure of critical piping

3.2.2 Defining Safety Management

Systems (SMS) The Safety Management System provides a HSE plan to ensure that the overall objectives for the management of all hazards and hazardous events (including those identified in Section 3.2.1 above and corresponding to fires and explosions) are achieved. This overall management process is outlined in the UKOOA guidelines [14] and COMAH regulations [105]. Based on the above documents, the management process for fires and explosions hazards is achieved through safety management systems consisting of the following steps:

• Identification of the hazards; • Assessment of hazards; • Reduction of hazards based on inherently

safe design principles, to reach a design solution where risk is ALARP; and

• Control and mitigate against residual risk.

The Safety Management System is based on managing hazards and hazard effects throughout the life cycle of the project, from conceptual design through commissioning and operation to decommissioning.

Blast and fire strategy for piping The Safety Management System will have a Health Safety and Environment (HS & E) plan, which will include a blast and fire strategy, which in turn will include a plan for blast and fire protection strategy for piping, which should address:

• Identifying safety critical piping; • Identifying responsibility of various

disciplines and interaction between them; • Developing a procedure for design and

protection of piping against fires and explosions;

• Developing a procedure for the design and protection of piping supports against fires and explosions; and



• Control and mitigation measures for the protection of piping and piping supports against fires and explosions.

3.2.3 Defining performance standards Performance standards provide a system of indicators that allow measurement of the successful (or otherwise) achievement of the goals.

The HSC Prevention of Fire and Explosion, and Emergency Response on Offshore Installations (PFEER)[16] regulations defines performance standard as:

A statement which can be expressed in qualitative or quantitative terms, of the performance required of a system, item of equipment , persons or procedure, and which is used as the basis for managing the hazard, e.g. planning, measuring, control or audit – through the life cycle of the installation.

The UKOOA guidelines on fire and explosion hazard management [14] proposes a hierarchy of performance standards:

• High level performance standards, which are applied to the installation as a whole and to major systems that constitute the installation; and

• Low level performance standards, which are applied to measure the performance of sub-systems, whose performance may affect the high level systems that are measured using high level performance standards.

In accordance with the goals described in Section 3.2.1 above, the level and number of performance standards should reflect the potential risk of the system whose performance they are intended to measure.

High level performance standards These performance standards are meant to measure the goals for the safety of the installation and relate to the overall risk to the persons on the installation. Fires and explosions and their effect on the topsides facilities and onshore process plants) will contribute to some of this risk.

For example, the performance of the overall blowdown system, and the fire and explosion water deluge system, will form part of the major

systems whose performance is to be measured. Examples of such high level performance requirements, related to piping systems include:

• The blow down system shall remain operational for an explosion event corresponding to a 10-4 return period;

• That water deluge systems will be operational in case of explosion or fire corresponding to a 10-4 return period; and

• That piping containing flammable, poisonous or explosive material will not fail, or will fail in a safe manner that will not lead to an unacceptable escalation of an initial event.

Low level performance standards Using the same example of the performance of the overall blowdown system, lower level performance standards should measure the performance of the elements and subsystems that comprise the blow down and fire and explosion deluge systems, and in turn contribute to successfully achieving the goals reflected by the high level performance standards.

They also apply to systems that could fail in a second area due to far field effects of an explosion in a first area, thereby leading to escalation of the incident in the first area to the second area.

Hierarchy of performance standards The PFEER regulations [16] use performance standards and acceptance criteria in an interchangeable manner. However, HS (G) 65 [15] indicates that performance standards can also be used to cover safety goals, and acceptance standards for various tasks in the hazard management process. In addition to defining levels of performance standards in terms of low level or high level (as described above) it is possible to adopt a slightly different approach, as described below (see Figure 3.4 [32]):

• Risk-based performance standards which are quantitative and specify levels of individual risk, fatal accident rate, or similar quantities which have to be satisfied;

• Scenario-based performance standards which can be either qualitative or quantitative, and which set an overall target or objective for the installation or part



thereof and complement the risk based standard; and

• Systems-based performance standards that specify quantitatively a minimum level of performance that must be demonstrated by personnel equipment, or design features under specified conditions.

The scenario-based and the system-based performance standards are more difficult to determine. However five contributing factors to the establishment of these standards have been identified:

Functionality; Reliability; Availability; Survivability; and Dependability.

Figure 3.4 Hierarchy of performance

standards [32]

3.2.4 Link between QRA and engineering acceptance criteria

OTO 2000 051 [8] gives a general overview of the relationship between Dimensioning Accidental Loads, engineering acceptance criteria and Quantitative Risk Assessment (QRA). Dimensioning Accidental Loads (DAL) are loads for those accidental events where the associated risks exceed the risk tolerability criteria. Therefore, the designed facility should successfully resist the DAL.

The requirements for successful resistance of a facility to DAL are expressed in the form of the low level performance standards (or specifications). Using the example of the

blowdown system in the previous Section, the performance standard should state that the blowdown piping should survive and remain functional during a postulated fire and/or explosion event corresponding to a 10000 year return period. In terms of engineering acceptance criteria this means that applied stress in the piping is not to exceed a defined allowable stress throughout the duration of an explosion and fire.

QRA uses “rule sets” to consider the effects of accidental loads on process facilities onshore and offshore. Rule sets for the survivability and functionality of pressure systems affected by explosions and / or fires should realistically predict the behaviour of the systems. Unrealistic rule sets contribute to the uncertainties in QRA and may lead to over-conservatism or inadequate design.

The following example illustrates the approach which includes a measure of design iteration for the system:

• A rule set in the QRA states that a blowdown system will perform a successful depressurisation during the explosions and fires resulting from events of a return frequency of 10-4 per year or higher;

• For the blowdown piping, this rule set assumes that the piping is functional, available on demand and survives the explosions and fires (performance standard);

• The piping is designed for the estimated 10000 year return period explosions and fires to meet the engineering acceptance criteria for stress, deformation, temperature, etc;

• QRA is completed and outcome events with the return period of 10000 years or higher are determined;

• DAL is determined for the blowdown system for the 10000 year return period outcome events; and

• The blowdown piping is re-calculated and re-designed, if necessary, for the DAL.

Another link between risk assessments and engineering criteria lies in the notion of safety margins in a strain-stress representation of reliability where load factors in structural codes were developed based on risk principles.



3.2.5 Guidance on ALARP decisions The concepts underlying ALARP are given in the HSE Reducing Risks, Protecting People (R2P2) document [18] and in the Guidance on ALARP for Offshore Division Inspectors [19]. Some of the main points are summarised below:

• Risk criteria and tolerability: The HSE framework for tolerability of risk shows three regions (see Figure 3.5):

A region of high risk, where the risk is unacceptable regardless of the level of benefit associated with the activity;

A region of intermediate risk, where the risk can be tolerated if it can be proved that there is gross disproportion between risk and further risk reduction, and if there is a system in place to ensure that risks are periodically reviewed to examine whether further controls are appropriate; and

A region of low risk where no additional measures are necessary except maintaining usual precautions.

• In the ALARP context the duty holder is required to take into account the individual risk and the societal risk (risk of multiple fatalities) bearing in mind that other aspects of societal concern have already been reflected in the regulatory regime in which the duty holder is operating;

• The HSE guidance indicates that it is good practice (but not enforceable) to apply the principles of prevention as a hierarchy;

• Good design principles aim to eliminate a hazard in preference to controlling the hazard, and controlling the hazard in preference to providing personal protective equipment;

• A holistic approach is important in order to ensure that risk-reduction measures adopted to address one hazard do not disproportionately increase risk due to other hazards, nor compromise the associated risk control measures; and

• It is expected that new installation would not give rise to residual risk levels greater than those achieved by the best of existing practice.

Unacceptable Region

Tolerable Region

Broadly Acceptable

RegionIn

crre

asin

g in

divi

dual

risk

s and

soci

etal

con

cern

s

Figure 3.5 Risk regions and ALARP [18]

Throughout the life cycle of the installation from the conceptual stage to the operation and decommissioning stages, risks should be assessed and risk reduction measures should be carried out if the risks are not ALARP. During the conceptual stage a wide variety of risk reduction measures are available, including prevention and elimination, while at later stages in the life cycle the majority of risk reduction measures available would fall under the control and mitigation categories. Figure 3.6 and Figure 3.7 show the various categories of available risk reduction measures and their variation from least to most preferred, where it can be seen that inherent safety, as discussed in Section 3.2.7, is the most preferred risk reduction measure.



Figure 3.6 The application of risk reduction measures at various stages [33]



Figure 3.7 Types of available risk reduction measures [34]

3.2.6 Demonstration of gross disproportion

Section 3.2.5 identified the reduction of hazards to as Low As reasonably Practicable (ALARP) as one of the main goals of the hazard management philosophy. Section 3.1.6 reviewed available guidance for HSE inspectors on how to assess whether the ALARP demonstration is acceptable. It was mentioned that the risks will be reduced to ALARP if it can be demonstrated for all available risk reduction measures that the benefit of implementing each hazard reduction measure is grossly disproportionate to the cost associated with implementing it.

To be able to make such a demonstration, the following issues must be addressed:

• The factors that should be included in determining the Cost of a particular Hazard Reduction Measure (CHRM); and

• The factors that should be included in determining the Value of a particular Hazard Reduction Measure (VHRM).

The measure that is used to demonstrate that the benefit of implementing each hazard reduction measure is grossly disproportionate to the cost associated with implementing it is the Ratio for gross level of disproportion: CHRM/VHRM.

Once this ratio exceeds a certain limiting value, it may be argued that the benefit of implementing a hazard reduction measure is grossly disproportionate to the cost associated with implementing it. The HSE documents [18, 19 and 24] provide guidance on the calculation of CHRM and VHRM. These documents also provide guidance on the limiting values of the ratio CHRM/VHRM at which the benefit of implementing a particular hazard reduction measure becomes grossly disproportionate to the cost associated with implementing it. Based on the above documents, Figure 3.8 shows that ratios of gross level of disproportion vary with the level of risk. For high risks, the ratio for gross level of disproportion may be as high as 10 and the cost reduction measure will have to be implemented. For low risks, near the negligible risk boundary, values of the ratio for gross level of disproportion as low as 1 are sufficient to demonstrate that costs are grossly disproportionate to benefit. However, it should be recognized that these ratios, together with the factors that are included in calculating costs (CHRM) and benefits (VHRM), are often subject to negotiation between the duty holder and the regulator.



The Individual Risk (IR) in Figure 3.6 expresses the probability per year of fatality for 1 individual. It is sometimes also termed as Individual Risk per Annum (IRPA). The IR depends on the location of the individual at a

given time and his / her scope of work. It should be noted that the values of IR in the Figure are examples only. They are normally determined by the operator at the onset of the project as a part of risk criteria.

Figure 3.8 Cost benefit ratios to demonstrate gross disproportionality

The risks associated with offshore and onshore facilities are often expressed in the form of Fatal Accident Rate (FAR). FAR is defined as the potential number of fatalities in a group of people exposed for a specific exposure time to the activity in question. Generally, the FAR is expressed as a probability of fatality per 100 million exposure hours for a given activity. The 100 million exposure hours is to represent the number of hours at work in 1000 working lifetimes.

3.2.7 Inherent Safety Figure 3.9 [35] shows the main stages in the life cycle of an offshore installation (Concept Design, Front End Engineering Design (FEED), Detailed Design, Construction, and Operation Maintenance and Control). At early stages the information quality is low while the influence on design is high. At later stages the quality of information is high, but the influence on design is low. The Safety Case Regulations do not provide a clear definition of inherent safety, however it

provides several examples of how it should be applied, including:

• Substituting less hazardous for more hazardous processes;

• Avoiding undue complexity in the design; • Allowance for human factors or control

systems which reduce the risk of human error; and

• The design of vessels and pipelines to minimise the effect of sources of deterioration, to reduce stress concentrations, and facilitate inspection after construction and during operation.

In regard to the design of piping systems, separation/segregation of units/modules (by barriers or distancing) is an inherently safe way of preventing escalation of hazardous event in a module or unit to adjacent modules or units (domino effects) which would otherwise enlarge the final consequence of the initiating event.



Figure 3.9 Outline of Life-Cycle [35]

OTO 98 148 [20], OTO 98 149 [21], OTO 98 150[22], OTO 98 151 [23] identify two alternative definitions of inherent safety:

• The first definition is related to design process- i.e. any activity which is carried out during the design to make the installation less vulnerable to environmental and man-made hazards. The effect of inherent design in this context is to reduce the likelihood of a hazard occurring, to reduce its consequence if it occurs, or in some other manner to reduce the risk associated with the hazard. Inherent in this context implies that vulnerability to hazards does not increase significantly over time, for example it is not dependent on repairs; and

• The second definition is not tied to the design stage, and can involve steps taken at the construction, operation or alteration stages. However it is restricted in the sense that it refers to actions that may be carried out to prevent a hazard from occurring. In this context, reducing the consequences of an incident once it has occurred is not as inherently safe as taking measures to reduce the likelihood of an incident occurring.

Between those two definitions the reports identify many other hybrid definitions which are linked both to elimination/prevention and to design. OTO 98 148 [20] reviewed 220 hazard management measures, and identified a trend where inherent avoidance is better than procedural mitigation. The hazard management measures were categorised under sub-topics, corresponding to various stages in the life cycle of the installation, as shown in Table 3.5 for a

variety of hazards, and in Table 3.6 which provide hazard reduction measures for piping systems.

Table 3.5 Hazard management measures, for a variety of hazards [20]

Applied in design Installation and operation

Other

• Robust and redundant design

• Layout and separation

• Design for blast pressure

• Use of appropriate design standards and work practices

• Use of competent design engineers / contractors

• Reduce manning • Reduce hazard • Reduce offshore

activity • Design for

people • Design for

weather tolerance

• Design for seismic activity

• Passive fire protection

• Procedural measures to avoid ship collision

• Fire and gas detection and fighting systems

• Devices to prevent dropped objects and collisions

• Procedural controls

• Inspection methods and philosophies

• Cathodic protection

• Floating vessels

• Control of modifications

• Emergency measures



Table 3.6 Fire and explosion hazard reduction measures for piping

Conceptual Phase Minimise inventory of combustible material Process and compress gas onshore to reduce processing risks offshore Avoid high energy systems Select less hazardous materials Hold materials in a form, or under conditions, to render them non / less hazardous Use less hazardous materials Include systems for flaring Provide a TR on an adjacent bridge linked structure Promote permit to work culture Build accommodation platform separate from production platform Separate personnel from process hazards Improved means of escape

FEED Phase Maximise ventilation including use of blow out panels Optimise deck height and equipment density Select and design blast equipment to withstand blast pressure Ensure adequate supply and maintenance of deluge systems Divide the inventory to reduce the amount with a potential to ignite Reduce the number of flanges and increase flange rating for critical piping Design for maximum pressure Ensure critical pipelines do not rupture when subjected to blast induced pressure Use welding rather than bolts Use welded pipework instead of flanged connections Use compact flanges Adopt pipe routes with shielding and running behind beams Adopt pipe routes avoiding vent areas Optimise location and level of pipe racks Optimise blast and fire protection (blast and fire walls)

Design Phase Improve layout of equipment and minimize congestion Design blast walls for high over pressures Minimise penetrations through blast walls, and provide seals where required to avoid transferring blast loading to penetrating services Deluge system feeders and their manual bypass lines to be protected by providing PFP cladding or coating to piping and where necessary to supports, or by extended deluge cover. Specifically for deluge feeders and bypass lines that skirt the separator area. Main blowdown header to be protected from fire in high hazard areas. Specifically relevant to header in vicinity of gas export metering package All emergency shutdown valves which are recognised to be critical in isolating major inventories will satisfy fail to safe condition Optimise fire protection coatings Optimise equipment fixings and piping supports

Construction Phase Provide and incorporate quality assurance of construction of piping and piping flanges and supports into overall safety management system Reduce as much as possible welding on site

Operational Phase Permit to work culture Provide operational training for all critical tasks and incorporate within safety management system Provide regular maintenance and inspection within safety management system



3.2.8 Interaction between fire and explosion hazard reduction measures for piping systems

The nature of interaction between explosion and fire will depend on whether an explosion precedes a fire or whether it occurs during a fire. Issues to be considered include:

• Effect of explosion on active systems that require an action to be taken before the system can become effective. A common form of an active system is the firewater deluge system, which comprises a firewater ringmain that distributes water to a network of small diameter pipes and nozzles, with valves controlling the flow of water. Therefore, the firewater ringmain should be designed and assessed for its ability to resist explosion loading, drag wind, differential support movement and fire loading;

• Effect of explosions on passive systems that do not require external activation to become effective. Such systems include firewalls and fire protection coatings, where the latter limit the temperature rise of underlying material by providing insulation or absorbing heat. Issues to be considered include strain compatibility and bonding of passive fire protection;

• Effect of explosions on control systems such as blowdown. A blowdown system must survive an initial explosion and/or fire, and remain functional after the explosion / fire to effect the emergency depressurisation. Furthermore, blowdown valve and piping must be designed for low temperatures, which occur due the rapidly expanding stream of gas through the blowdown valve;

• Effect of fires on explosion resistance. For explosion occurring during or after a fire, the effect of fire on the explosion resistance can be significant, as yield strength and Young’s modulus are reduced at elevated temperature. These should be taken into account, especially as explosion resistant design involves utilising the plastic deformation capacity of members, which may be reduced because of the fire. In general it is not common practice to assess explosion resistance during a fire, as in such effects vapour cloud explosions may be ignored. Pressure vessels have to exhibit adequate resistance against fire to prevent

their rupture or BLEVE. A pressure vessel rupture would release large flammable inventory and may generate projectiles. The latter may damage fire / blast walls and result in large escalation that may develop into a domino effect;

• Fire protection coatings can be optimised to reduce their extent of application and thickness. This reduces the aerodynamic diameter and in turn the explosion drag and overpressure. Also, less coating needs to be inspected and removed for inspection and maintenance of the pipework and equipment, thus reducing the plant downtime and exposure of personnel to hazards;

• Compact flanges are less prone to leaks when properly installed, which can reduce the frequency of accidental releases. They also are smaller than API flanges, thus reducing the congestion and blockage, which result in the reduction of explosion overpressure; and

• Safety conflicts, where different safety consideration for fires and explosions are contradictory and where some ‘trading-off’ of benefits is necessary. Since most fatalities usually occur due to smoke inhalation under fire conditions, fire considerations may be more in focus than explosion. Some common conflicts are presented in Table 3.7.

Table 3.7 Safety conflicts between fire and explosion considerations

Conflict Description Segregation versus openness

Segregation of a platform using solid fire walls could result in an increase in overpressure when an explosion occurs

Ventilation versus weather protection

Ventilation is desirable because it reduces gas cloud size and therefore overpressure due to explosions, however it can lead to the spread of smoke due to fires and flames outside modules

Deluge versus ignition

If the deluge system is to be triggered on gas detection, care should be taken not to increase the probability of ignition

3.2.9 Zoning and sectionalisation This Technical Note uses the terms of “zones” in conjunction with area classification, “fire and explosion area” associated with potential escalation, and “sectionalisation” as a



subdivision of plant into isolateable process segments. These terms and their relationship are explained below.

Area classification The aim of area classification is to minimise the likelihood of ignition of those releases that inevitably occur from time to time in the operation of facilities handling flammable liquids and vapours. The approach is to reduce to an acceptable minimum level the probability of coincidence of a flammable atmosphere and an electrical or other source of ignition [36].

It should be noted that area classification cannot guard against the ignition of major releases of flammable materials under catastrophic failure of plant, e.g. the rupture of a pressure vessel or pipeline, or the cold failure of a tank which, in properly run facilities, has a very low probability of occurrence. The incidence of such releases must be kept within acceptable limits by correct design, construction, maintenance and operation of facilities.

Area classification is the assessed division of a facility into hazardous areas and non-hazardous areas, and the subdivision of the hazardous areas into zones. The Institute of Petroleum classify three hazardous area zones [36]:

Zone 0 is that part of a hazardous area in which a flammable atmosphere is continuously present or present for long periods.

Zone 1 is that part of a hazardous area in which a flammable atmosphere is likely to occur in normal operation.

Zone 2 is that part of a hazardous area in which a flammable atmosphere is not likely to occur in normal operation and, if it occurs, will exist only for a short period.

Fire and explosion area Fire and / or explosion area is defined as an area limited by firewalls, blastwalls, combined fire- and blastwalls or space, which aim is to prevent the escalation from one fire / explosion area into another.

In onshore plants it is the space between process units or plant plots that provide separation and minimises the potential of escalation from one unit to another, and between plots.

Sectionalisation It is common practice that every process plant is sub-divided into isolateable process segments for the purpose of inspection, maintenance and depressurisation. A process segment denotes all piping and equipment within one depressurisation volume. The ESD and PSD valves connected to the segment define the battery limit of the depressurisation volume. A single pressurised vessel, storage or transportation tank, etc., can also be a process segment. During depressurisation, the pressure in the process segment is reduced through the BDV and its depressurisation orifice.

Combined effects of zones, areas and process sections Area classification zones do not extend across boundaries of fire and explosion areas. This is given by spacing in onshore plants, and by fire- and blastwalls or spacing on offshore installations.

Process plants handling hydrocarbons are normally classified as Zone 1 or Zone 2 and there is electrical equipment located within these zones designed with appropriate protection.

The area around the flare or vent tip is normally Zone 0. Equipment that may cause ignition based on its operation is not located in Zone 0.

Pump motors are one group of obvious sources of ignition of accidental leaks from a process section in a hydrocarbon processing facility. However, the ignition probability in risk analysis takes a conservative approach and assumes failure of the pump motor protection in Zone 1 or Zone 2.

Initial explosion and / or fire from a process segment should not escalate beyond the fire / blastwalls, i.e. it should be contained within the fire / explosion area around the section. This puts requirements on the explosion and fire protection of emergency shutdown and blowdown valves. They have to remain functional until such time as the risk of escalation is reduced to an acceptable level.

3.3 Determination of criticality levels for piping

It would be impractical and costly to design all piping systems to resist all fire and explosion



scenarios. Piping systems should therefore be screened and classified to criticality level ratings and the fire and explosion resistance designed accordingly. Two criticality level ratings are proposed, level 1 and level 2; piping systems of criticality level 1 are more critical than those corresponding to criticality level 2. It is the intention that other pipes (pipes not belonging to either criticality level) not be subject to assessment for fire and explosion loads.

3.3.1 The determination of criticality levels

Criticality Level 1 Piping Piping systems that may cause the loss of a main safety function should be of criticality level 1. These systems should be designed such that accidental loads with a return period greater or equal to 10000 years shall not cause such loss [37].

The following main safety functions must be maintained in the event of an accident situation on a permanently manned offshore facility:

1. Preventing escalation of accident situations so that personnel outside the immediate vicinity of the accident are not injured;

2. Maintaining the main load carrying capacity in load bearing structures until the facility has been evacuated;

3. Protecting rooms of significance for harm limitation due to accidental events, so that they are operative until the facility has been evacuated;

4. Protecting the facility’s safe areas so that they remain intact until the facility has been evacuated; and

5. Maintaining at least one escape route from every area where personnel may be sheltering until evacuation to the facility’s safe areas and rescue personnel has been completed.

The above criteria apply in general terms to onshore plant but additional objectives would be:

1. To limit the spread of damage in a site;

2. To prevent escalation to tank farms and storage areas for hazardous materials; and

3. to prevent spread of the leaked hazardous gasses to people / facilities outside the site boundary.

Installations may be designed for accidental loads of return periods other than 10000 years, but the selected value should reflect the operator’s risk criteria.

The approach would call for carrying out the QRA prior to the determination of the criticality level and pipework design for accidental loads. This may cause problems with the design schedule, as the QRA is a relatively lengthy process. This situation can be resolved by carrying out a coarse hazard evaluation and use the results together with past experience to screen for piping systems of criticality level 1. During further development of the design and / or completion of the QRA it may become necessary to adjust the selection of criticality systems. The criticality levels outlined in Sections 3.3.2 and 3.3.3 are suggested as guidance based on past experience.

Criticality Level 2 Piping Systems of criticality level 2 are designed to withstand a lesser event with a shorter return period. Their objective is to reduce the escalation potential and / or limit the extent of damage.

Criticality Level of Piping Supports The criticality level of piping supports is the same as the criticality level of the piping they support. Supports of critical piping should be designed more rigorously than supports for non-critical piping.

3.3.2 Criticality Level 1 piping Critical piping systems level 1 are typically:

• The pressure relief/blowdown system itself, including pressure relief/blowdown piping, flare header, knock-out drum and all branches to pressure systems that are to be blown-down;

• Risers upstream of the emergency shutdown valves (ESDVs), ESDVs themselves and their actuators;

• Hydrocarbon lines, such as large lines with systems containing large quantities of hazardous fluids that would discharge in the event of failure. This category may include



wellhead piping, flow lines and link lines to main storage tank farms;

• High-pressure gas lines associated with gas compressors, where the inventory potentially released is large. The key here is the sectionalising philosophy for the gas compression plant. It should be recognised that a suitable sectioning philosophy might allow these lines to be downgraded to criticality level 2; and

• In special circumstances, there may be other lines containing significant inventories of poisonous gases or chemicals that are potentially hazardous to the people or the environment.

3.3.3 Criticality level 2 piping Critical piping systems level 2 are typically:

• Pipes whose failure would contribute to the size of a fire. The idea is not to design for a less severe explosion scenario than the dimensioning scenario. Damage of criticality level 2 equipment will be severe close to the heart of the explosion but the spatial extent of the damage in a given explosion/fire scenario will be reduced;

• Piping whose failure could cause an additional leak and a second explosion ignited by the first fire. The objective is to prevent escalation of an explosion in one module or plot/unit to an adjacent one. Its implementation would reduce the allowable distancing between individual plot/units;

• Piping that would not otherwise be provided with any explosion resistance at all, eg pipes within buildings and modules that are not subject to meteorological wind and therefore have no external load design requirement other than gravity loads;

• Unprotected pipe and flanged connections, especially connections to high-pressure vessels, whose failure could lead to jet propulsion loads on the pipe or vessel components either side of a break. These components could then cause escalation to nearby areas; and

• Pipes within existing modules and plot/units that become critical as a result of the addition of new equipment or plot units nearby.

3.3.4 Explosion load corresponding to Criticality Level 2

Given the range of objectives in allocating criticality 2 to systems, there cannot be one load criteria to relate design loads for criticality 2 to criticality 1, hence it is not possible to make criticality 1 and 2 design loads generally analogous to a DLB and SLB criteria for structure design.

Given that a large number of lines could come into the criticality 2 Category, design must be by a practical method and the target strength level must be an achievable one. Given sufficient distancing or provision of blast barriers the design quasistatic drag load could be quite low in some circumstances, e.g. 0.05bar drag.

For extensions to existing installations a criticality 2 rating for equipment in the existing parts of the installation would have to be set at a realistically low level, given the constraints of the retrofit situation.

The resistance of pipes to criticality 2 loading may be assessed and validated by simpler means (e.g. Section 6.3.1).



4. PLANT LAYOUT

4.1 Overall layout aspects Safety input to layout is a perquisite for minimising headline design pressure in explosions and for minimising escalation potential for fires. This is as applicable for floating units and onshore plant as it is for fixed offshore installations. The layout for accidental events influences the overall platform concept and in the case of onshore plant, the overall plot layout and the land-take for the proposed plant.

The Interim Guidance Notes [2], CMR Gas explosion handbook [4] and ISO 13702 [26] deal with fundamental layout issues in relation to minimising or exacerbating explosion pressures.

This Section discusses measures specific to reducing piping design loads and minimising escalation potential. The use of specific types of flanges is addressed in Section 4.4, as it affects the aspects of local layout. As different sizes of different types of flanges may affect the overall layout.

4.2 Offshore installations 4.2.1 Fixed Offshore Platforms Explosion In general, a long and narrow topsides is better from the explosion standpoint as it leads to lower probable loading, and potentially allows for more distancing between hazardous and non-

hazardous zones. The feasibility of this approach depends on the choice of substructure and this aspect should be considered when selecting the substructure concept. Once the substructure concept is chosen, the only overall layout option that can fundamentally affect design pressures and drag loads is the selection of module width to height ratio: a key design decision is the setting of the module deck-to-deck height and locating the main pipe racks.

The objective in increasing the heights of modules is that good ventilation is easier to achieve, modules will be shorter or narrower for the same quantity of contained equipment and the length of the flame path is shortened, reducing pressures. Furthermore, the pipe runs in the zone of flame acceleration (secondary and tertiary pipe racks in Figure 4.2) are predominantly aligned with the product outflow, contributing much less to turbulence: this further reduces headline pressure.

This policy, however, leads to long and narrow topsides and is not a viable option unless detail aspects are considered early on in the design when substructure and topsides structure layout options are being considered.

Where the use of blastwalls is contemplated, the module deck-to-deck height should be limited to 10 - 11m so that blastwalls are not higher than 10m (the availability of bending machines with a capacity of more than 10m is very limited).



Figure 4.1 Fixed structure layout [38]

N

Equipment

primary pipe rack

secondary pipe rack

tertiary pipe rack

Figure 4.2 Typical piping layout on a fixed structure [6]

Fire Due to the density of equipment, pipework and structures, all fires inside process modules will be obstructed. Flames from fires originating from pipework runs along decks outside modules may be unobstructed if the leak points outboard over the sea.

Firewalls and plated decks and floors should prevent the penetration of a jet flame, unless the jet flame is abrasive due to its sand contents and

high momentum. In cases where firewalls or plated decks do not form a complete enclosure around the fire source, large flames impacting on the walls / decks / floors may exhibit a “curl-over” effect around them. The curling-over part of the flame will have the intensity of a background load and it may threaten personnel and equipment on the side being protected.

The flame of large ventilation-controlled fire in closed modules tends to fill the module. This

TR utilities Process

LQ

Cabins Module

Switch gear room Emergency generator Firewater pumps, Cooling & heating pumps

Separators

Flare drum Export metering Export pumps

Compressors

Main generators

Gridlines



together with the heat re-radiation back to the flame further supports the combustion process, which results in very high heat loads of 400kW/m2 and complete engulfment of equipment, pipework and structures within the module.

Explosion and fire On fixed offshore platforms it is possible to locate safety routes and control lines at different levels, and to provide isolation between levels with solid decks (as can be seen from Figure 4.1 [38].

On fixed offshore installations, especially those built as integrated decks, main pipe racks are often located at the perimeter of process areas, as can be seen from Figure 4.2 [6]. In some respects this is not desirable, as it can block the free ventilation, increasing probable cloud sizes and the size of credible explosions. Furthermore, by reducing vent area it increases explosion pressures. On the other hand, such a location may reduce the escalation potential as the main (hydrocarbon containing) pipes may be separated from the source of the fire by a firewall. The decision which solution would be better to use should be based on risk analysis.

4.2.2 FPSOs Explosion The process equipment on these units is normally located on a raised deck or on Pre-Assembled Units (PAUs) above the main hull deck. This forms a barrier between the tanker piping, which is a lower explosion hazard, and the process piping. It also reduces the extent of green water due to storms on the deck and leads to more space for process equipment and a more manageable construction methodology. In addition, it constitutes a barrier between the congested process areas, where explosion of ignited gas clouds pose a large risk, and the and the cargo deck areas, where there are large hydrocarbon inventories.

On large floating units the process area can be very extensive and congested and may outgrow a single deck. Pipe support design then becomes quite difficult because, unlike a fixed platform, there are no readily available strong decks to which the higher level pipe racks can be fixed. The problem is not how to design the supports of the equipment and pipes, but rather it is one of deciding what sort of above deck space frame

should be constructed to tie the pipe supports back to.

Another aspect of these units is that the deck is large and usually open, so that an explosion will have effects over a wide area, as can be seen from Figure 4.3 [39]. As discussed previously, on fixed platforms it is possible to locate safety routes and control lines at different levels, and to provide isolation between levels with solid decks (as can be seen from Figure 4.1 [38]). With floating unit design this is only feasible where the process deck is solid: a problem occurs if the turret is between the process area and the TR.

Main pipe racks usually run parallel to the dominant blast wind direction, ie longitudinally (primary rack in Figure 4.3). In practice, transverse wind effects are least at the centre line of the unit where the pipe rack is usually located, as can be seen from Figure 4.4 [6]. Pipe racks have a significant number of transverse beams so longitudinal forces will be significant. Where the pipe rack is diverted to run transverse to the axis of the unit, severe sideways wind loads can be expected.

Secondary and tertiary pipe racks are subject to the most severe loading in the unit N-S direction (see Figure 4.4). They have to be supported from an overhead grillage of beams, which in turn are supported by portal frames. When the quantity of tertiary racking is included the overall N-S force becomes large and it is not easy to see how the force can be resisted unless a brace frame concept is adopted for the whole superstructure. The forces in the bracing will be almost entirely explosion wind driven: hence the overall stability of the pipe racking will depend on the stability of the frame.

Section 4.4 provides a discussion on how ductile construction for pipe racks can be designed to yield low quasi-static design explosion loads, and therefore increase allowable design drag loads.

On FPSOs the pipe rack is usually located on the centreline and the drag will be predominantly longitudinal. When the pipe rack is located at the ship’s side, rather than on its centreline, the transverse drag forces can be large especially on a large unit: some are up to 60m wide (see Figure 4.5).



Living Quarter

AirTreatment

ControlRoom

Power generation

Electrical &InstrumentationBuilding

Seawaterlifting

Deareationcolumn

FiltrationDesulphatation

Injectionpumps

Desalting

Production watertreatment

Oil / gas / waterseparation

Oil manifold

Flare

Dehydration & gascompression

Oil metering& offloading

Figure 4.3 Typical detail of FPSO layout without turret [39]

Figure 4.4 Typical piping layout on an FPSO [6]

N

Equipment

primary pipe rack

secondary pipe rack

tertiary pipe rack



Figure 4.5 FPSOs centreline rack has lower loads than side rack

Fire Fires originating within the central area of a floating unit will be obstructed by equipment, pipework and structures. Due to the narrow and long shape of the units, the probability that the main pipe racks and pipes are engulfed is high for medium and large horizontal or near-horizontal jet fires.

Main escape routes are normally located at the longitudinal sides of the deck and protected by a firewall. It should be noted that a jet flame impinging on the firewall will curl over its top edge and may render a substantial length of the escape route unavailable.

Enclosed modules are not normally used on floating units and the use of fire-rated blast walls is usually limited to one or two protecting the living quarters and also segregating some other parts of the facility. Again, there will be a curl-over effect over the top and edges of the wall impinged by a fire that may affect the personnel, equipment and pipework located on its protected side close to the wall edges.

The likelihood of escalation of medium and large fires to neighbouring process segments will be higher than that on fixed facilities. This is because there may only be limited physical compartmentalisation, such as on two-deck floaters, where process piping on the upper deck is segregated by its plated floor from the storage piping on the hull deck.

Plated decks in PAUs will offer some protection, however, flames will curl around the deck edges

and may affect equipment, pipework and structures on the cold side.

Escape tunnels Some floating units have escape routes enclosed in tunnels and located at the longitudinal sides of the unit. Such escape tunnels provide improved protection if they are correctly designed for both explosion and fire.

FPSO turrets The routing and design of pipes leading from fluid swivel stacks on FPSO turrets into the main piping system is a specialist area. This requires a pipe bridge rack that has very large cyclic movements due to hull flexing in waves (which imposes severe fatigue loading). Catering for these effects means a long span articulated pipe rack structure, which can be difficult to design when it is necessary to consider blast. These issues are fundamental to the topsides layout and need preliminary study at the beginning of the process design.

The vertical pipes in internal turrets on FPSOs see almost no drag loading, only high levels of field pressure. With these it is important not to have bulky pipe supports or horizontal expansion bends in the top half of the turret where they can increase flame acceleration to the restricted top vent, increasing field pressures.

4.3 Onshore plants Onshore plants generally comprise plots separated by open space. A typical plot layout is shown diagrammatically in Figure 4.6. An

process pallets

short flame path to pipe rack

long flame path to pipe rack



individual plot usually represents a complete integrated process system.

Plots are divided by smaller gaps into sub-plots or units whose size will be dictated by the nature of the process being undertaken within it. Storage of large hazardous inventories is invariably in tank farms located on plots that are separate from the processing plots and sometimes at considerable distance from them. Long pipe racks link the tank farms to the process plots.

In principal, tank farms (where the large inventories are) and process plots (where the congestion and most of the high pressure gas and leak sources are located) are separated so that explosions and / or fires in the process plots cannot affect the tank farms and fires in the tank farms cannot cause explosions / fires in the process plots (inherently safe design principles).

4.3.1 Critical piping The problem comes with processes which have large inventories within them, such as LNG plant and processes that require large residence times for chemical reactions. In these cases an explosion can lead to a severe fire and it can therefore be necessary to design piping for the explosion effects so that unacceptable escalation is prevented.

Two other concerns are the prevention of leaks of poisonous gasses to nearby populated areas and cases where the distance between process plots is insufficient to prevent a large explosion in one plot having a large far-field effect in an adjacent plot and a consequent risk of domino effects. In such cases it can be advantageous to design piping and equipment to withstand drag and blast loads.

A special aspect of LNG plants is that the low temperature vapour, is mostly buoyant at temperatures above –100°C. However, once mixed with air, it can reach temperatures much higher than –100°C at which point the LNG becomes heavier than air and therefore in the explodable range. This means that LNG leaks in

one plot / area can lead to explodable vapour clouds enveloping other plot/units in much the same way as heavier than air gasses do.

In onshore process plant, the rack often forms the spine of the plant, with the process plant located some distance to each side, with long horizontal legs into each process sub-plot. These lateral pipes can be quite long, with valves in them, and would be the critical sections for explosion design, with both a transverse and vertical component of drag.

Other pipes interconnect with larger vessels and process buildings which can be subject to movement in blast. These pipes may not have much load applied directly to them but need to be made of a ductile form of construction so that they can accommodate the movement of the equipment without further leaks occurring (where such leaks are linked to significant inventories). Furthermore they may need fire protection, especially to the flanges.

4.3.2 Effect of insulation on explosion loads

Another aspect of onshore plant is that many of the pipes are insulated, for heat conservation and sound reduction. Insulation increases the outside diameter of pipes and the drag loads on them in proportion. It also increases turbulence and explosion pressure. Insulation must therefore be allowed for, both in the explosion pressures analysis and in the response analysis. This is often not easy to do because the insulation is not shown explicitly on drawings and in CAD models, merely as a code letter in the line numbering scheme.

On the other hand, thermal insulation may provide protection against fire. This is conditional upon the insulation surviving and remaining functional after the initial explosion and during the fire.



Figure 4.6 Typical layout for a plot or part-plot of an onshore plant

4.3.3 Plant modifications and brownfield sites

For modifications and extensions to existing plant the scope for modification of existing components is likely to be limited. The evaluation of the new plant components should take into account additional potential releases and the consequences of fires and explosions in the new plant components on existing components. The re-evaluation of existing components should also be carried out and should take into account the consequences of fires and explosions from new/modified components. The minimum goal in such projects should be to minimise the incremental risk and consequence for the plant as a whole and, when opportunities present themselves, use should be made of the modification project as a means of reducing overall risk.

The number of sources of potential additional releases can be minimised by the use of fully welded connections. However, this should not be considered solely on its own merits. For offshore installations, welding in situ increases the probability of ignition and requires platform shutdown in most cases. The use of welding habitats may be considered, although one should learn from both good and bad experience with this solution. Another possibility is to pre-assemble process units onshore to minimise welding offshore. Also, the ease of maintenance offshore should be considered, where the exposure of personnel to occupational and process risks should be minimised.

Plant modifications or extensions may influence the emergency response plan. It should be confirmed that essential safety systems and personnel act in a harmonised manner after the modifications / extensions are implemented.

Explosion risk Far field effects In the context of this Section far-field explosion drag is defined as drag acting on piping and equipment located outside the boundaries of the module / unit where the fire / explosion accident took place. In general it is far-field drag that can escalate the effects of an initial explosion. Therefore, the detailed objectives of such projects are met by balancing the provision of adequate distancing and increasing the design drag resistance of the plant.

To avoid spread of accident effects beyond the boundaries of a module / unit, distancing can be employed and achieved by suitable siting of new plant. Figure 4.7 shows four different options during plant extension. Figure 4.7i shows the new extension location at a distance sufficiently far such that the effects of an explosion will not spread from the existing part of the plant (a) to the new extension (b) or vice versa. Alternatively overall risk reduction can be achieved by relocating a set of existing functions from the original plant into the new plant extension thereby creating an inactive or non-hazardous process zone as a barrier within the existing plant (Figure 4.7iii). Where such barrier

Open roadway (fire break)

spine rack

Long flame path drag predominantly parallel to rack



zones could lead to ignition and flame acceleration within them (thereby bridging the adjacent zones – see Figure 4.7iv) the inactive plant should be removed.

Physical barriers (fire-rated blastwalls) erected within the plant can be used to prevent escalation to neighbouring areas but their downside is that they tend to decrease natural ventilation. The ventilation reduction increases the probability of flammable cloud-forming and ignition. Furthermore, the explosion will become confined, leading to a pressure increase. Of

potentially particular interest is that physical barriers can be very effective in open plant for separating plant where heavier than air gas is being processed from plant where lighter than air gas is being processed. The walls restrict the spread of heavy clouds along the ground and deflect the spread of lighter than air clouds over the wall and above congested equipment zones. The confining effect of blastwalls can be reduced where an empty space is provided between the wall and the hazardous plant (both sides of the wall if hazardous plant is on both sides).

a b

a b

(i) hazard reduction by distancing of new plant - equipment (a) and (b) do not affect each other

(ii) distancing of new plant not adequate - equipment (a) and (b) do affect each other

Nonhazardous

a b

a b

Leak

Ignition

(iii) hazard reduction by relocation of hazardous process zone

(iv) non hazardous zone with ignition source

Figure 4.7 Hazard reduction measures during plant modifications



.

Empt

y sp

ace

Ignition

Heavierthan air

Ignition

Blast wall

Blast wall

Figure 4.8 Plan view of open plant with blast walls

Sacrificial blast walls (e.g. explosion relief walls) can also perform this function but it should be recognized that in severe explosions the explosion relief panels will (however well designed) inevitably detach, or the support structure be picked up by the explosion loading, causing a projectile hazard for down-stream plant and persons. (see FABIG TN2 Section 5 for assessment procedure [40]).

Reduction of risk to life in the extended plant can sometimes be achieved by building underground escape tunnels (onshore plant) or by routing escape routes along a different deck level (offshore plant).

Fire risk The characteristics of potential fires may change after the modifications are implemented following changes, as for example, in fluid

composition, pressure and / or water cut. This may increase or reduce the flammability of the fluids in case of accidental release.

The use of the existing facilities and fire resistance should be always taken into consideration. The capacity of the essential safety systems should be confirmed for the situations after the modifications are implemented.

The capacity of depressurisation systems for upset as well as accidental fire conditions should be checked in an early stage of the design of the modifications. This especially true for older designs where depressurisation systems may have been designed in accordance with API RP 521 for unrealistically low heat fluxes. An excessive demand for flaring capacity may be solved by staggered blowdown.



It is very unlikely that fires related to the modifications will penetrate existing firewalls, however, they may be of such intensities and duration that make the firewall ratings invalid resulting in excessive temperature rise on the cold side of the firewall.

Floor plating will generally prevent flame penetration, unless the erosion effect of the flame is extremely high due to large volume of sand in the fluid leaking at a very high pressure over a long period of time. However, if the flame is sufficiently long it will curl around the plating and may affect equipment on the other side.

Based on recent calculations and experiments flanges impinged by flame will lose their tightness and become source of additional leaks within a few minutes after the start of the fire unless they have fire protection coatings. This will escalate the initial fire into additional fires, however, the pressurised pipework and equipment will depressurise quicker (as a result of the additional leaks) resulting in shorter flames and shorter fire duration. If this limited escalation can be controlled within a module or a plant unit, it will help to prevent escalation to a neighbouring module or unit and can reduce the chance of domino effect and cataclysmic fire throughout the plant.

Passive fire protection coatings can be optimised to provide adequate protection at a minimum extent and thickness. This reduced thickness results in a decrease in the value of drag and overpressure from the explosion. Also the risk of corrosion beneath the coatings is reduced, which reduces the leak frequency.

4.4 Projectile risk Projectiles can arise in the initial explosion or in a subsequent fire. In the first case there are two types [5]:

• Primary projectiles i.e. fragments of barriers which are initially accelerated by differential explosion pressures across the barrier and subsequently by explosion wind action on barrier particles; and

• Secondary projectiles which are items picked up and driven by the explosion wind.

In addition, projectiles can occur during a fire when a pressure-containing item bursts due to

heating and weakening in a fire. In the initial explosion projectiles can occur due to the failure of a large flange connected to a pressure containing item. The projectile is accelerated by the rocket-motor effect of compressed material exiting the breach. The breach might be a tear rupture in a vessel or pipe or a failed flange connection. Very high kinetic energies and large flight distances can be achieved by projectiles. The exiting material may be on fire and liquid drop-out during the flight path can spread fire to areas below the trajectory.

Risk reduction on aspects of overall layout involves:

• Increased distancing, though how to judge an adequate value would be hard to define;

• Use of energy absorbing (ductile) blastwalls; and

• Use of massive walls such as thick concrete or earth filled crib walls.

At the local level risk reduction consists of one or more of the following:

• Improving pipe and vessel support arrangements;

• Reducing likelihood of failure in fire by application of PFP, especially to flanges; and

• Improving blowdown rates and replacing/supplementing pressure relief with blowdown

• Using orifice plates at flanges so that flow rates at break points beyond the orifice plates are restricted. This requires that the weak point is not the flanged connection where the orifice plate is located.

Given the difficulty of defining the energy and flight distance of projectiles quantitatively reliance is likely to be placed on local risk reduction measures supplemented by the blanket protection afforded by providing adequate distancing and protection from far field explosion pressures as described elsewhere in this document.

4.5 Local layout aspects While this Section is geared to offshore platforms, some aspects may also be applicable to onshore plants.



Equipment location Local layout optimisation requires the hazardous areas (and some parts of adjacent areas) to be zoned according to peak drag pressure and direction. Location of equipment should take account of likely drag loading in an explosion and the sensitivity of the equipment and its criticality rating.

Differential support movement Pipes aligned along the flow direction cause less turbulence (less flame acceleration) and have lower sideways bending loads. The piping runs which are parallel to the predominant drag flow will have a larger interval between pipe supports and consequently more flexibility. The advantage of this aspect should be taken by using these pipe spans for accommodating differential support movement.

Vessels Vessels that cross the outflow of explosion products will have enhanced drag flow above and below where critical piping systems might be located, e.g. blow-down equipment.

A vessel in the vicinity of a blastwall and parallel to it will not see this phenomenon because the blast wall prevents flow across the vessel. On the other hand, pressure piling could happen under the vessel, locally increasing blastwall design load. This might be more easily catered for with an enhanced blastwall, especially if the blastwall has ductile deflection capacity (pressure pile-ups are short-duration pressure increases), but this risk needs to be known prior to ordering the blastwall.

Nozzles Nozzles on some vessels and gas compressor barrels are often 1 – 2 sizes smaller than the pipes that connect to them: they represent therefore a weak point in the piping system that can limit capacity, with the failure mode being potentially a full-bore release (large cloud). It is important to limit drag loading to conservative values in such cases by good layout design (location and orientation of compressors). It is not by providing extra support that the problem due to thermal, vibration and fatigue constraints in the pipes may be solved. Process sectionalization and fire strategy is an important issue that might enable the criticality of these pipes to be dropped to level 2, with a lower

design load and a more amenable set of constraints on piping design.

Flare headers Flare headers, large lines containing flammable fluids and firewater headers are usually criticality level 1. Many parts of them are welded but flanged connections at the PSVs and BDVs are inevitable. In other situations it is possible to isolate these connections from blast loading by locating a strong fixed support nearby. With high-level blow-down lines, the lack of a nearby solid deck structure makes it difficult to solve the problem satisfactorily at the level of the pipe support arrangement. Pipe routing from the PSV is important and aligning the first leg of the pipe with the dominant drag direction will go a long way to reducing the problem.

Flare headers may have to slope continuously from the gas entry point to the knock-out drum, so it is not possible to shelter them behind the perimeter girders everywhere. In the next Sections, some guidance is given on pipe rack design and on some aspects of flare pipes that might limit their blast capacity. These aspects can affect the piping design and need to be tackled at the piping specification (P&ID) stage.

Flanges and fittings The number of flanges in a petrochemical plant may be up to several thousands for large plants. They take space and have a considerable weight. Compact flanges [41] were developed in the beginning of the nineties, and these offer advantages in some areas in comparison with ordinary API flanges.

The estimated leakage rates for compact flanges are lower in comparison with standard offshore and standard quality flanges [42]. A summary is given in Table 4.1. Frequency values for piping in the Table are given only for comparison.



Table 4.1 Estimated leakage frequencies per year for flanges

Description Small / Medium Leakage

Large Leakage

Compact flange 2.5 x 10-5 1.1 x 10-5 Standard offshore flange

4.1 x 10-4 3.0 x 10-5

Standard quality flange

10 x 10-4 10 x 10-5

12m piping including one girth weld

1.7 x 10-4 1.6 x 10-5

Average value for 1m piping with one girth weld per 12m

1.4 x 10-5

1.3 x 10-6

Compact flanges are smaller than API flanges for the same diameter, and rating of the pipe [42]. This results in a lower aero-dynamic diameter, generating lower explosion drag load. Also, because of the smaller size of the compact flanges, the congestion and blockage is reduced, resulting in lower explosion overpressure.

Flanges, due to their numbers in a plant generate a considerable load on the structure, especially on offshore facilities. The lower weight of compact flanges leads to a lower required strength of the supporting structure, further resulting in smaller structural members, and less required fire protection by weight and application area.

It has been shown by calculations and experiments that unprotected flanges will lose their tightness within a few minutes after the start of a fire [43, 37]. As the leaks from high-pressure hydrocarbon pipework would lead to large fires, and the large fires have high escalation potential when affecting high-pressure pipes, it is desirable to fully weld high-pressure pipework. This may be combined with the use of compact flanges as their leak frequency is lower than that of ordinary flanges.

When deciding about the extent of flanges and fully welded pipework it should be borne in mind that while fully welded pipework may offer low leak frequencies, the use of flanged connections will reduce the time required for inspection, maintenance and modifications. This would in turn reduce the exposure time of maintenance personnel to risks.

It should also be noted that flanges, piping tees, and tappings located in areas of high bending moments may restrict the blast capacity of

piping systems. These limitations on capacity may not be recognised when carrying out code-checking piping stress analysis to ASME B31.3. However, these limitations on capacity will need to be accounted for when SDOF and MDOF response methods are applied. This shows the benefit of carrying out a preliminary multi-discipline blast assessment before locating and orientating branch lines.

Complexity of support configurations An important issue is that complex solutions to piping stress problems (addition of supports or increasing their complexity) can increase the amount of tertiary steel added, and therefore the degree of congestion. This, in turn, will increase turbulence in an explosion and therefore explosion pressure.

This is a particular consideration with long pipe racks in onshore plant and FPSOs, where it is important to make the rack assembly as smooth and streamlined as possible for longitudinal (drag) flow. If the congestion for longitudinal flame acceleration is kept below a certain limit, the flame will not accelerate significantly along the pipe rack.

Fire break zones In the event of a really large gas cloud covering several plots or PAUs along a pipe rack separated by open gaps but linked by the pipe rack, the flame front will slow down at each plot interval (“fire break”) rather than continue to speed up along the pipe rack.

In onshore plant it has not so far been the practice to design piping for drag, and as a consequence many pipe rack structures have only small amounts of pipe support and tertiary steel and hence probably conform already to this “smooth” concept. It would be a disadvantage if bringing in a requirement to design for drag were responded to by spoiling this attribute.

Expansion loops Another layout point to be considered is that multiple expansion loops, which may either dangle down beneath the pipe rack or be horizontal within it, should not be placed in the “fire break” zones (although one or two would probably be acceptable in many cases). It is actually feasible to avoid intermediate expansion loops by having the right sort of sliding supports and sufficient lateral leg length on the pipe



where it comes into and leaves the pipe rack. These lateral legs can; however, be a design problem, especially the valves at the equipment end. Also, the tie down methods for the pipes need careful consideration, given the large expansion movements.

The pipe expansion design also needs to accommodate the thermal expansion of the pipework due to fire.

4.6 Non-conductive materials and minimising electrostatic sparking risk

Non-conductive materials such as plastics, GRP and Epoxy-based intumescent PFP can under certain circumstances become charged with static electricity. If the charge is discharged to earth in an explosive atmosphere and the discharge energy is sufficient, an explosion or fire would result.

The conditions for charging, the presence of a suitable discharge point to earth, and the presence of an explosive atmosphere must all be present for an ignition and explosion to occur. In principal the probabilities of all the conditions being in place are low and the risk is generally low. However, there are circumstances and design aspects that can favour electrostatic sparking and these need to be addressed by reference to CENELEC TC31 388 and TR 504-04 “Code of Practice for the avoidance of hazards due to static electricity and Technical Report”.

The condition of the non-conductive material at the time of discharge affects ignition risk. Many non-conductive materials absorb water from the atmosphere, in the long term this can reduce discharge risk and energy. Conversely, a cold vapour leak in the vicinity may lower the instantaneous humidity and adversely affect discharge risk [44].

In CENELEC, Zone 2 hazardous areas are exempt from treatment, because of the low

probability of gaseous atmosphere being present. However, in Risk-Based design the risk of occurrence of explosive mixtures in Zone 2 areas is acknowledged and some form of assessment may be required to ensure that the particular aspect of the design of the facility does not contribute significantly to ignition risk.

In some ways this approach is consistent with the adoption be some operators of Zone 1 classified electrical equipment in Zone 2 areas.

Factors which increase electrostatic charge or spark ignition risk are:

1. Non-conducting surfaces or unearthed conducting components;

2. Some aspects of size and shape of the conductive surface;

3. Unearthed conductive components and conductive paints on non-conducting surfaces (increases spark energy);

4. Low ambient humidity (prior to or at the time of a gas leak);

5. High momentum droplet sprays during gas leaks; and

6. High speed flow of low conductivity fluids in plastic pipes: charging of the outside surface of the non-conducting material is then possible.

Non-conductive materials can be checked for sparking risk by coupon test. Inherent safety can be achieved by using materials which cannot produce electrostatic sparks of sufficient energy to ignite the gas or by avoiding the use of non-conductive materials. GRP and Intumescent Epoxies can usually be made inherently conducting by the addition of Carbon black to the resin.

Metal and other conducting components mounted on non conducting supports should always be earthed.



PART C DESIGN



5. EXPLOSION LOADS

5.1 Parameters affecting explosion loading

The parameters affecting explosion loading are [2, 4, 5]:

• The process segment where the leaks come from;

• Leak location and direction; • Ventilation; • The size and concentration of the ignited

cloud of flammable gas; • Shape of module where explosion takes

place; • Degree of congestion; • Ignition location; and • Time of ignition.

5.1.1 Process segment The composition of the leaking gas, pressure and temperature, release size and the volume of gas available for leaking will all vary between process segments. Representative process segments may need to be established to keep the explosion load analysis manageable.

5.1.2 Leak location and direction The location of the leak within an area or module and its direction will have influence on the size and shape of the gas cloud.

5.1.3 Ventilation Ventilation depends on the wind direction.

Ventilation on one side or on two adjacent sides of an offshore module is not as effective as ventilation on two opposite sides (this, again, would decrease the maximum distance a flame can travel before reaching open air).

For land-based plant ventilation will be influenced by air flow around large items of equipment, process buildings and office buildings.

5.1.4 Cloud size and concentration The size of the ignited gas cloud will influence the magnitude of explosion. In addition, the gas

concentration in the air/gas mixture cloud will not be constant.

For the cloud area of a stoichiometric mixture: The larger the cloud area of stoichiometric mixture, the higher the resulting overpressure.

For the cloud area that has concentration lower or higher than stoichiometric: The closer the cloud concentration will be to the stoichiometric mixture, the higher the resulting overpressure.

5.1.5 Shape of module Modules (not topsides, see Section 4.2.1) closer to the shape of a cube are better than those closer to the shape of a rectangle, with one or more sides significantly bigger than the others. This is because the further a flame has to travel before it reaches open air, the more it will increase in strength.

Similarly for land-based plant or process building, the longer the flame has to travel before it reaches open space, the higher the explosion overpressure.

5.1.6 Congestion The density of equipment, pipework and structures (congestion) and whether or not they are orientated such that they block the path that the explosion flame travels (blockage) will affect the build-up of explosion overpressure.

The more an offshore module is congested, the more turbulence the flame is likely to undergo as it travels towards the exit, and therefore the higher the generated loading. The orientation of the obstacles can also play a significant part in increasing or controlling the loading. Obstacles arranged in series in the path of flame propagation will have a more detrimental effect on the loading than obstacles arranged in parallel (see Table 5.1). In addition, obstacles blocking part of the vent location will also cause an increase in the loading.

The higher the congestion and / or blockage of a land-based plant or process building, the higher the generated overpressure. For example, for a plant / process building in a shape of a long rectangle, large diameter pipework and vessels should be preferably orientated with their longitudinal axes in the direction of the rectangle length.



5.1.7 Ignition location Location of ignition relative to the vent is also very important; the nearer the ignition location is to the vent, or the end of the plant / process building, the less distance the flame has to travel, and in general the less the loading.

5.1.8 Time of ignition The later the ignition after the start of the leak, the higher the chance is that the gas cloud will become flammable.

Also, a delayed ignition may cause the explosion overpressure be higher than in case of an early ignition, as the flammable cloud may be larger with time.

These have all been discussed in greater detail in explosion handbooks [2, 4 and 5]. Guidance for their modelling may be found in Annex G of NORSOK Standard Z-013 [45].

Table 5.1 Effect of various parameters on explosion loading [2, 4, 5]

Parameter Bad Good Shape

L>> W or L>>H

L = W = H Ventilation - cube

One side or two adjacent sides Two opposite sides or 3 or 4 sides Congestion – repeated obstacles

Repeated obstacles Improved layout - Obstacle groups separated by large open breaks>10m wide

Congestion vs. venting

Equipment blocking vent area Improved layout - Vent paths and exits clear of equipment and buildings

Ignition location

Ignition at opposite end of vent location Ignition in vicinity of vent location - except for modules open at one end only, where pressure piling becomes a risk

Fuel-air Fuel-air



5.1.9 Loading direction The overpressure and drag loading, and the differential displacement loading, can have two lateral components (X and Y) and one vertical component (Z). In most instances drag and overpressure will be dominant in one direction only. In addition, the drag and overpressure loading will have a positive and negative phases as can be seen in Figure 5.1 and Figure 5.2, respectively.

Rebound following a drag impulse can also occur even where there is no flow reversal. This means that in a typical quasi-static analysis, piping and supports will need to be designed for explosion loads in the direction of drag loading and in the opposite direction. Guidance on rebound dynamic load factors for some common impulse shapes is given in FABIG Technical Note 4 [46].

N

Equipment

primary pipe rack

secondary pipe rack

tertiary pipe rack

Figure 5.1 Drag and overpressure loading – positive phase [46 and 6]

N

Equipment

primary pipe rack

secondary pipe rack

tertiary pipe rack

Figure 5.2 Drag and overpressure loading - rebound (negative) phase [46 and 6]



5.2 Methods used for determining the explosion loading

Explosion loading components acting on piping systems include drag and overpressure components of the loading. The differential displacement (Section 5.2.6) at the piping supports is also considered a loading acting on the piping system even if it is a response-effect of the topside structure.

The general characteristics of drag loading are:

• Drag loading is directional and depends upon where the cloud is and where the vent or end of the plant / process building is located. As credible clouds do not fill the whole module / plant / process building, drag at any given location will vary in direction according to cloud location;

• drag loading gets worse as you move away from the centre of the cloud; and

• it is usually worst near the vent exits.

This Section starts by giving theoretical background for the determination of drag loading on pipes (Section 5.2.1). Next various methods with increasing order of accuracy for determining the drag loading on pipes are discussed:

• Nominal values of drag and overpressure (Section 5.2.2);

• Drag pressure based on field overpressure (Section 5.2.3);

• Drag force based on Baker Cd values (Section 5.2.4); and

• Direct load measurement (Section 5.2.5).

5.2.1 Theoretical considerations The force on a rigid, stationary object in a moving fluid is a result of the spatial and temporal variation in the fluid pressure and flow around that object. For example, a pressure wave passing the object will result in a time varying pressure gradient across it. In addition, the flow induced by the pressure wave will be deflected by the object, inducing further pressure gradients (due to normal stresses) and frictional forces (due to tangential stresses).

The pressure distribution around a cylinder gives a time varying force in the stream-wise and the

crosswise direction. The aerodynamic force is given by the integrated normal and shear stresses over the surface of the obstacle. The following equation shows a phenomenological splitting of the aerodynamic force:

( )

HEMDP

dh

FFtVUt

Ut

mVtUtUACF

++∂∂

+

∂∂

++=

)()(

)()()(21

ρ

ρ

The first term is given by quasi-static form drag depending on obstacle shape, surface roughness, fluid density and Reynolds Number. The second term is the inertia force proportional to the acceleration. This term consists of the buoyancy term of the obstacle in the accelerated flow field and an added mass term from the integrated normal stress 180 degrees out of phase with the bulk flow motion. The third term is a combustion contribution originating from the change of mass per unit volume in a generalised version of Newton’s law. The fourth term is the differential pressure, which is a function of the Mach Number (U/c), where c is the velocity of sound. The last term is the hydro-elastic term.

Overpressure calculation has been discussed in detail in many publications. As far as piping is concerned, it is important to establish whether the calculation method will account for both overpressure and drag, or whether drag loading will have to be determined separately.

All of the most important components of drag and overpressure loading are accounted for in the Computational Fluid Dynamics (CFD) models. The analyst has the option to report drag impulses (which only accounts for form drag) or drag force by the direct load measurement (DLM) method (see Section 5.2.5), which accounts for the other drag load elements (last four terms in above equation).

Design loads for small obstacles (diameter < 0.3m) may be computed using the drag impulse method. Values for large obstacles should be computed using the direct load measurement method, which takes into account the other drag load components. These other force components are small and out of phase with the form drag component when diameter < 0.3m. However, they can become significant when small diameter components are in close proximity to each other.



5.2.2 Nominal values of drag and overpressure

It is possible to have nominal values of drag and overpressure based on various parameters affecting the explosion and drag (such as ventilation, congestion, layout, etc…). Values of drag pressure between 0.15bar to 0.3bar are often quoted in the literature; however, these values have not been arrived at through a systematic approach that considered the various factors affecting drag. While these values may be used in the absence of any other data, currently there is no qualitative way to take into account how these nominal values may decrease or increase in view of some simple changes being introduced to the installation (e.g. increase congestion, increase ventilation, etc..).

5.2.3 Drag pressure based on field overpressure

Another simple approach is to use a relationship between drag and field pressure, and use the former for loading on the piping system.

Figure 5.3, based on a recent FABIG Newsletter Article [47], below shows such a generic overpressure relationship for offshore modules and FPSOs. Firstly this Figure shows that the relation is not linear but has a trend towards drag pressure being a larger proportion of overpressure as the overpressure value increases, with the ratio being 0.2 at 1bar and 0.35 at 2bar.

Figure 5.4 shows the trend of overpressure and drag as the cloud size increases (up to 80% fill) for a tunnel-shaped process module 35m x 19m x 7m. This confirms the same trend, with a peak drag pressure of a little under 0.2bar for 1bar peak field pressure and 0.45bar for 2bar field pressure.

It is important to recognise that reducing the headline overpressure reduces the design drag pressure. If the headline 10000 year return period overpressure for this module was 1bar, corresponding to 30% cloud fill, the design drag pressure would be slightly below 0.2bar.

In this Figure “TOP 4” value means the mean of highest 4 measuring points out of about 50 in the module. Many parts of the module could have lower design pressure and analysis of results can support a “zoning philosophy” which would allow some areas to be classified as low drag areas where one could elect to locate sensitive equipment such as gas compressors.

If the headline overpressure were 2bar, corresponding to 47% fill (in this example), the design drag pressure would be 0.45bar. This indicates the importance of good ventilation for reducing the loading.

Experience on offshore facilities has shown that designing piping to withstand 0.2bar drag is generally regarded as feasible but 0.5bar is much more difficult and in some cases impractical. Example 1 in the Appendix is based on 0.2bar and the piping arrangement shown, though in principal quite robust, may be difficult to upgrade to 0.5bar quasi-static resistance.

Drag Load Correlation for Pipework

0

50

100

150

200

250

300

0 1 2 3 4 5

Explosion Overpressure (bar)

Drag

Loa

d (k

N/m

2)

Figure 5.3 Drag / overpressure

relationship



Figure 5.4 Example of sensitivity of pressure and drag loading to cloud size

The UKOOA update [33] of the Interim Guidance Notes [2] state that in the absence of any detailed information it is possible to use a ratio of drag load to overpressure load equal to 1/3; where the overpressure load at a particular location would be determined based on the Ductility Level Blast (DLB) load case.

Zoning for drag As mentioned above, the overpressure and drag will vary significantly through the module. In general the highest drag pressures occur near the vent openings (fixed offshore platforms and plant confined within buildings) and lowest near the module centre. This difference can be captured in design by applying a zoning philosophy with two levels of design drag pressure. The first level of design drag pressure relates to “TOP4” drag impulses in high-drag zones and a second level of design drag pressure relates to “TOP4” impulses in remaining zones. In FPSOs and onshore plant the same philosophy can probably be applied but it may not be clear how to define the demarcation line between the zones. Furthermore, in extensively congested plant there may not be any low drag areas within the congested plant.

5.2.4 Drag force based on Baker Cd values

Drag loading The drag force, a combination of the skin friction (tangential stresses) and the form drag (normal stresses), is commonly written as follows:

vvACF dd ρ21

=

where ρ is the fluid density, A is the maximum cross sectional area of the object in a plane normal to v, Cd is the drag coefficient and v is the large scale fluid velocity ignoring spatial fluctuations in the vicinity of the object.

Note also that in a turbulent flow the velocity term is actually a time-mean value, albeit over time-scales less than those of the large scale variation in the flow. This means that variations on turbulent time-scales are not included in this analysis. These include such effects as vortex shedding, in which vortices form alternately on one side and then another of the object. (The potential results of vortex shedding have been demonstrated in disasters such as the collapse of the First Tacoma Narrows bridge, which may partly be attributed to wind excited vortex shedding.)



The drag coefficient Cd depends on the shape of the obstacle and on the Reynolds number of the flow. The drag coefficient is a function of the Reynolds number and the local Mach number of the flow. The expression for the Reynolds number is:

μ

ρ D V Re =

Where: ρ is the fluid density V is the free-stream fluid velocity D is the pipe diameter, μ is the dynamic fluid viscosity The expression for the Mach number is: M = V / C

Where: V is the free-stream fluid velocity C is the speed of sound in the medium of propagation

The nature of Cd has been studied extensively for various standard shaped objects in steady flow. Its functional dependence on the Reynolds number is summarised in Figure 5.5 [48], for flow transverse to a cylinder. For low levels of

turbulence the skin friction dominates. At higher levels of turbulence, eddies form behind the object and the form drag dominates, as a result of which the drag coefficient levels out. This plateau corresponds to the commonly quoted values (e.g. see TNO Green book [49]) such as 1.2 for flow transverse to a cylinder and 0.82 for axial flow past a cylinder. The sharp drop around Re=106 is the so-called drag crisis, where a sudden onset of boundary layer turbulence, although increasing skin friction, moves the separation point further back on the object reducing the form drag.

Both the Reynolds and the Mach numbers are functions of the time-varying flow conditions and so the drag coefficient is also time varying. Figure 5.6 shows the dependence of the drag coefficient on the Mach (Hoerner, 1965) number [50]. If the Reynolds number is greater than 4 x 105, the flow is considered turbulent and the Mach curve shown in Figure 5.6 should be used. For lower Reynolds numbers, the Reynolds curve shown in Figure 5.5 should be used.

Figure 5.5 Cd for a cylinder, as a function of Reynolds number (Schilting, 1960 [48])



Figure 5.6 Variation of drag coefficient for a cylinder with Mach number [50]

Figure 5.7 DLM Method: determining pressure differential on large obstacles

The following subsections discuss the applicability of the drag loading method to various piping size and obstacle shapes.

Large obstacles (diameter >2m): The drag method (on its own) is not suitable for determining explosion loading on large obstacles. The DLM Method [6] may be used for large obstacles and obstacle groups. Loading is like the inertia component of wave loading in

marine design. In CFD programmes, one monitor point is located upstream of the obstacle and one downstream of it (see Figure 5.7). Within the CFD programme the differential pressure (ΔP) across the obstacle or obstacle group is plotted against time. The force is determined from the differential pressure using the expression Force = ΔP x windage area x PDF where PDF = Pressure Distribution Factor which needs to be selected on the basis of the



shape of the obstacle group. For cylinders one might use PDF = 2/π, which presupposes a sinusoidal pressure distribution around the obstacle. For a rectangular object one can probably use 1.0.

Small obstacles (diameter < 0.3m): The drag method may be used for smaller obstacles: “Drag” pressure = 0.5 ρ v2. In CFD programmes the local density of gas/burnt products, ρ, varies widely according to gas temperature and pressure. It is therefore useful to examine the CFD output in terms of “drag” directly rather than v. As drag is directional, one can specify that drag should be reported

individually in the two lateral and one vertical directions.

The design load for a pipe is = 0.5 Cd ρ v2 D where D is pipe diameter and Cd is the drag coefficient.

For diameters up to 0.3m the loading components other than drag are relatively small and these can, in general, be ignored.

Baker’s Cd values [51] (Table 5.2) may be over-conservative for cylinders subject to explosion wind but this may enable drag method to be used for pipes up to 0.5m diameter (see OTO 99 046 [6]).



Table 5.2 Drag coefficients, Cd, for various shapes [Baker 51]

Sketch Shape CD Flow

Circular cylinder (long rod), side on

1.20

Sphere 0.47

Flow

Rod, end-on 0.82

Flowor

Disc, face-on 1.17

Flow

Cube, face-on 1.05

Flow

Cube, edge-on 0.80

Flow

Long rectangular member, face-on

2.05

Flow

Long rectangular member, edge-on

1.55

Flow

Narrow strip, face-on 1.98



Intermediate obstacles (0.3m < diameter <2m): The DLM method is difficult to apply for diameters below 2m unless control volumes (element mesh size) in the CFD analysis are reduced. Control volumes (usually cubic) of 1m or 0.5m are typical but for obstacles down to 0.3m in size, control volumes of 0.1m or smaller are required. Figure 5.8 shows the use of the DLM method with standard size control volumes (0.5m to 1m) to determine pressure differential on intermediate obstacles (e.g. 1.0 m). Since the size of the obstacle does not correspond to the size of the control volume, it can be seen that the measurement does not correspond to the peak pressure differential across the obstacle. With obstacles in the diameter range 0.3m to 0.5m, (and possibly up to 1m) it is probably best to use the drag method and use Baker’s conservative drag coefficients: (CD = 1.2 for pipe).

Figure 5.8 DLM Method: determining

pressure differential on intermediate obstacles using standard size control volumes

Obstacle groups and pipe-racks These often comprise a mixture of large and small obstacles supported on a common structure. Both the time-history of forces acting on the common structure and the time-history acting on the individual obstacles (e.g. pipes and pipe supports) are required.

Consider an example of a 40m long, 15m high and 6m wide pipe-rack on an FPSO or an onshore chemical plant.

The load on the obstacle groups may be obtained using either the drag method or the direct load measurement method. However, the problem with the drag method is how to establish a drag windage area for the whole assembly taking into account the shielding effect of obstacles behind each other and flow acceleration effects of objects one above the other (horizontal cross flow).

The CFD programme actually finds the pressure drop across the pipe rack. The differential loading between each pair of opposing monitor points represents the applied loading on the section of the pipe rack that is located between the monitor points. Depending upon how many monitor points are used, one may or may not have to apply a PDF (pressure distribution factor). As the PDF is usually less than 1.0, it is usually conservative use a PDF value of 1.0 and have the monitor points approximately level with the centre of windage area or the level with the densest obstacle zone.

An alternative is to obtain the drag components at suitable locations within the pipe rack zone, determine shielding factors for proximity of pipes from a conventional wind code and sum up the total windage area of the rack as a “porosity”. The next step then is to set the drag force equal to the drag pressure for flow across the rack, multiplied by the rack side area and by the porosity factor.

Obstacles subject to far field blast In the far-field, explosion impulses tend to shock-up and behave like TNT blasts with a vertical shock front and a zero rise time. Experience of damage in terrorist explosions, such as Canary Wharf (1996) show that lamp-posts, parking meters and handrails and even the columns of buildings suffered no significant damage despite close proximity to the explosion epicentre. This is because the blast wave rapidly envelops the obstacle and equalises upstream and down stream [52]. With TNT type explosions the drag wave attenuates rapidly with distance from the explosion epicentre and therefore its far-field effects may be ignored. However, in refineries, or offshore deflagrations, the amount of burnt gas produced in the short time frame leads to a more significant wind pulse which spreads further afield and may need to be considered for piping design. This can have a knock-on consequence (domino effects) affecting plant outside the zone of the initial



deflagration. Unfortunately the Multi Energy Method commonly used for this hazard has limited ability to give the required design information, but CFD models can and have been used for this.

In the far field, the most common failure mode is “unit moves and pipes break” (Section 4.5, Table 3.1). In such instances, the shock loading causes movement of a large vessel (for example) and the pipes that connect to it do not have the ductile construction to withstand the movement without major leaks occurring. In such situations it is necessary to treat the movement of the equipment as an imposed simultaneous displacement of pipe supports in one part of the line relative to pipe supports in another part.

Pipes with large amplitude of vibration: aero-elastic enhancement of drag force. This phenomenon applies potentially to the following:

• pipes with very large span-diameter ratio, • pipes with low Young’s modulus materials,

e.g. GRP • flexible hoses.

The problem is that the explosion wind causes vortices to initiate in turn from one side of the obstacle and then the other, setting up a cross flow vibration. This cross-flow vibration causes the pipe to oscillate systematically across the flow into the accelerated flow region at the side of the pipe, increasing the time-averaged drag loading on the pipe. In effect it is increasing the apparent drag coefficient Cd.

There is also a vortex shedding induced cyclic load in line with the wind at twice the vortex shedding frequency. Tests of oscillating cylinders in steady water flow have shown an increase factor for drag is up to 3.0 when the cross-flow movement is a significant proportion of diameter [53].

In principle, the phenomenon only occurs when the pipe vibration frequency is close to the vortex shedding frequency. However there is a “lock–on” effect which causes the vortex shedding frequency to match the vibration frequency. This broadens the frequency band-width over which the effect can occur [53], and leads to an increase in the flow velocity range, which in turn may lead to an increase in the

duration of the phenomenon as a proportion of the drag impulse time.

Of course the drag impulse duration is short and the amount of movement of the pipe that is required for the phenomenon to occur is large. Hence, it can be shown that there are many situations where the phenomenon may safely be ignored.

Tank tests of articulated cylindrical towers in current with waves show that the phenomena are reduced, albeit not eliminated, in unsteady flow. Gas explosions involve turbulence and unsteady flow so the increase factor for Cd would be expected to be less, compared to steady flow tests.

Analytical methods for quantifying these effects are fraught with difficulties and it seems that explosion tests may be needed to quantify the effects. However, as far as current practice is concerned, a more qualitative approach is required.

In some cases it may be possible to shelter vulnerable pipes behind beams to keep drag loading off them. In other cases, where there is a risk of increased drag occurring it may be necessary to enhance the drag pressure for design purposes.

To assess the risk of the phenomenon occurring, the ratio of the drag impulse duration to the vibration frequency for the critical pipe vibration modes should be examined. Consider the example of a straight pipe in a pipe rack. The vortices may excite any one of a number of modes in the cross-flow direction, with a wide potential for lock-on. This would increase the drag for bending of the pipe in the flow direction (e.g. either the first or the second mode of vibration). Higher modes may be ruled out since they will undergo lower deflection for a given excitation force, while a certain amount of movement (as a proportion of pipe diameter) is required for the phenomenon to happen at all [53].

For example if the vibration amplitude (for the critical mode) was only a few percent of the pipe diameter (due to the high stiffness) it seems unlikely that the pipe would move far enough to create an enhanced drag.



Another factor is that if the pipe is a continuous tube, without fittings and penetrations it would not matter if it bent plastically under the loading (plastic energy absorption capacity). This of course would not be possible for GRP pipes, as they are not ductile.

The problem in all cases would be defining the design load for the pipe supports. In addition it is important to ensure that the adopted design method will provide supports with adequate compliance for potential pipe deflections.

For flexible pipes the key would be providing enough slack in the hose loop, thereby minimising vibration.

There could be a problem for small pipes located between adjacent large pipes in a dense pipe rack, where the flow up through the gaps between the adjacent pipes could be enhanced due to the overall blockage of the rack. A simple solution might be to tie the small pipes to adjacent large ones at points between supports. However, it would be important to determine carefully the pipe support design load for the larger pipes.

5.2.5 Direct Load Measurement based on CFD Analysis

Representative scenarios and return probability criteria for explosion loading In most offshore installations and hazardous onshore chemical plant it is feasible to have explosion scenarios which simply cannot be designed for. In such cases probabilistic analysis may be useful. The design explosion event will not then be the worst case but must have a low probability. For offshore this means 10-4 annual probability threshold (ISO 13702 [26] and ISO 19901.3[30]). For onshore plant public (individual) risk comes in as well and land use planning regulations [54] may apply.

The key factor here is that the size of the cloud, or at least the part of it that is within the flammability limits, is limited by ventilation and wind conditions.

In ISO 19901.3 [30] the 10-4 threshold applies for the aggregate of all the explosion events on the installation. For an installation with 4 modules this means the threshold for each module is, on average, 2.5 x 10-5, but only for events which breach the boundaries of the

module. For piping this means avoiding ruptures that lead to fires escalating beyond the boundaries of the module.

Fire containment strategy The fire containment strategy dictates the explosion design criteria for the piping design, and reduced criteria are possible if the overall tolerance of the intact system and the structure to fire is high enough to withstand serious fires in the failed parts. A good fire containment strategy will ensure that it does not matter if many pipes rupture.

For fixed platforms, fire containment strategy depends on blast rated firewalls, distancing process plant from people, and environmental risk (ISO 19901.3 [30]).

On onshore plants, the fire containment strategy depends on plot size, separation distance between plots, equipment density and magnitude of hazardous inventory.

On FPSOs with large amounts of hazardous inventory and where plots are very large with respect to the area of a FPSO, there is little room for space between plots. In addition, environmental risk is important. An important distinction between onshore plants and FPSOs is that the latter have a larger amount of fire water available per surface area. Nonetheless one should try and create spaces between plots in order to avoid having long runs of flame acceleration due to long runs of congestion. Blast walls and blast relief walls can also be used in this way, but the former tends to increase explosion pressure. These measures have beneficial effects on the probable size of gas clouds and, in the case of lighter than air gases are likely to be the main value. These considerations may also apply to fixed platforms with extensive process areas.

Arctic platforms are much more difficult due to poor ventilation, lack of escape possibilities and very high environmental risk. The requirements for protection from the weather for arctic platform will tend to increase confinement and lead to larger probable gas clouds in the event of a leak. Ice accretion on relief walls and louvers can restrict their opening and venting efficiency, and could therefore restrict operation and lead to production downtime in severe cases.



Exceedance curves Figure 5.9 shows a probabilistic presentation of data. Considering a 10-4 annual probability for this module and ignoring the effects of water deluge (the lower line), the design explosion pressure would be 1.45bar. Referring now to Figure 5.4 1.45bar corresponds to 40% cloud size (cloud size as a proportion of total module volume), so that scenarios with about 40% cloud fill would be representative. This means that for a scenario-based response analysis, one could select the 40% cloud fill cases for generating the design drag impulses from which to select the ‘TOP4’ values discussed in Section 3.4.3.

Another approach is to examine the pressure / cloud-size curves (similar to Figure 5.4) and

number the points according to explosion scenario number. The predominant scenarios around the 10-4 return period pressure (1 in 10000 probability return period) are the ones to select design scenario(s) from.

An alternative approach for determining the loading due to an explosion event is to generate drag loading exceedance curves, which may then be used with a particular return period. However, this approach is only used when a large amount of calculations are required. It is perhaps used when other more conservative methods generate very high explosion loading. Figure 5.10 [55] shows a typical overpressure exceedance curve. No equivalent drag exceedance curve is available in the open domain.

Figure 5.9 Example of a pressure exceedence curves

Figure 5.10 Overpressure exceedance curve [55]



5.2.6 Differential displacement In addition to drag and overpressure loading, a third component consisting of differential displacement at the piping supports should be accounted for. Figure 5.11 shows a segment of a typical high pressure piping collection system. It can be seen that the piping is restrained along various points of its length against movement in the x-, y- or z- direction. If these pipe supports are attached to the main structure, the pipes may be exposed to differential displacements of the structure as the latter is loaded by explosion overpressure.

Decks on which equipment items are mounted are subjected to differential pressure loads, which induce vertical (and sometimes horizontal) acceleration of the equipment. Deck accelerations principally cause vertical movements and forces in the equipment but, when the equipment is tall and not located at the centre of the deck, the deck inclines locally as well as moving vertically. This will induce horizontal movements and accelerations at the centre of gravity of the equipment.

Determining the amount of differential displacement at the piping supports requires an accurate estimation of the support stiffness and may include a degree of iteration as more information becomes available as the project progresses.

One possibility to determine differential displacement is to use past experience on

previous platforms or experience based on earlier analyses. Another approach is to select nominal differential displacement values that the piping system will be able to withstand. In addition, ductile construction methods may be used to ensure that the piping system will be able to undergo the differential displacement without rupturing. A discussion on ductile construction and local layout aspects is provided in Section 4.

A more accurate manner for determining deck displacements, which may be used in high risk cases, is to use the results of finite element analyses of a platform and deck models to determine deck pressure points and perhaps spatial plots of deck differential pressures. This data may then be used for calculating deck displacements and accelerations.

Both process and water deluge piping systems may be subjected to differential movements due to their particular support configurations. Figure 5.12 [47] shows a segment of a typical deluge system, which may be subjected to differential acceleration if it is supported at various points along its length. The model is subject to a dynamic load from the supporting structure, which is affected by explosion overpressure.

In many cases, differential deck acceleration can be a major contributor to loading acting on piping, more than loading due to either overpressure or drag.



Figure 5.11 Segment of a typical high pressure system [56]

x

y

z

Figure 5.12 Segment of a typical deluge piping system [47]

Elastic deck deflection For decks of offshore modules designed to respond elastically (e.g. when designing against the Strength Level Blast, SLB, criteria) the imposed support deflections will be relatively low but could still be significant when considering relative movement of the ceiling of the module to its floor. The relative (quasi-static) movements may be obtained from the

elastic structural analysis of module framework. Based on experience from structural analyses, span / deflection ratios for structural members are typically 1 / 130 or less.

Plastic deck deflection If plastic deformation is accounted for in determining resistance (e.g. when designing against the Ductility Level Blast, DLB, criteria)



deflections will reach or exceed span / 40. For a 15m deck span, the relative deflections of the floor to the ceiling could be 15000/40 x 2 = 750mm which may not be feasible for a piping system to accommodate, and the piping may lose its integrity.

In principal, critical piping connected to deck mounted equipment should all be anchored to the same deck as the equipment. In addition, the piping routing to and from such ceiling mounted equipment should be of a ductile form of

construction without bolted flanges and with suitable supports to prevent excessive bending moments being transmitted from the pipes to the equipment and flanged connections outside the designated movement zones.

Figure 5.13 shows a ductile form of construction where the pipe attached to the vessel is allowed to move between the pipe anchor and the pipe support. Furthermore, the pipe that is allowed to move is welded and does not have any flanges and bolts.

Pipe supportValve

Pipe anchor

Vesselsupport

StringerBlast wallbehind vessel

All welded pipe

Beam

C platformL

Figure 5.13 Pipe vessel attachment allowing for support movement

Variation of loading with time The drag and overpressure loading, as well as loads due to differential displacements, varies irregularly with time. In cases of design specifications and sensitivity studies, it is often useful if the loading can be simplified and linearised to a simple trapezoidal or triangular form. Several studies have focused on linearisation techniques including the gas explosion handbook, the Gas Explosion Model

Evaluation Protocol (MEGE [57]) and the Interim Guidance Notes [2].

The other alternative is to model the exact time variation of the loading, which may be more sensitive to differences between what is modelled and what is constructed. The full pressure-time history for each pair of opposing monitor points may be obtained from the results of CFD analyses. These may then be used with either SDOF or MDOF structural analysis methods.



6. DESIGN OF PIPING AGAINST EXPLOSIONS

6.1 Introduction This chapter considers the techniques that may be used to determine explosion loads and their effects on piping systems. Section 6.2 discusses the effects of explosions on pipework. Section 6.3 presents design flow charts for piping systems with three methods for piping stress analysis. It also gives guidelines for pipe support design, pipework design and ductile construction. Section 6.4 presents a discussion on acceptance criteria.

The design method is illustrated by an Example in Appendix 1 and this will be referred to as Example 1.

6.2 Effects of explosions on pipework

Possible effects of explosions on pipework include:

• Drag loading on pipes. • Dynamic effects on pipes. • Drag loads and dynamic effects on pipe

supports and bolted connections.

The dynamic effects may be due to:

• Inbound and rebound response of the pipe or its supports excited by a passing pressure wave.

• Excessive excitation due the natural frequency of the pipe or its supports being close to the dominant frequencies of the pressure impulse.

These effects on their own or in combination may result in:

• Brittle or ductile guillotine-type rupture of the pipe, leading to a large additional release of flammable material. The additional release is likely to ignite, forming a jet or pool fire, or a combination of both.

• Brittle or ductile fracture forming a crack in the pipe resulting in a medium or small release of flammable material and fire.

• Fracture of flange connections resulting in a full bore or a crack size release and fire.

• Excessive deformation of the pipe resulting in a medium or smaller fracture and fire.

If the release is located near the passing combustion products from the explosion, the released fluids would be immediately ignited and fire would ensue.

Excessive pipe deformation or a guillotine-type fracture may also result in a fracture and flammable gas release away from the explosion combustion products. In such a case, a flammable cloud may form resulting in a new vapour cloud explosion.

6.3 Design flow charts Figures 6.1 and 6.2 show a typical design flow chart. The explosion design activity would have a starting set of activities based on static analysis using ASME B31.3 and Caesar II or an equivalent code (Example 1). In this document, this is defined as Category 1 analysis. Enhancement options could be considered for pipes that do not pass normal ASME B31.3 criteria, e.g. enhanced allowable stresses for strain rate effects. For pipe materials, enhancement factors up to 1.2 might be applicable, depending on material selection. Lower enhancement factors would have to be used for high strength materials such as that used for bolts.

Three categories of piping stress analysis are possible, with increasing levels of accuracy and complexity. These are discussed in detail in following subsections. The selection of the method of analysis is subject to several factors including the principle of proportionality of risk (see Section 3.1.2) and additional factors discussed in Section 6.3.9.

The proposed way forward is that a piping system that passes Category 1 pipe stress analysis (ASME B31.3 code check) does not need to undergo a Category 2 piping stress analysis (ASME B31.3 together with SDOF analysis). Similarly a piping system that passes Category 2 piping stress analysis does not require a Category 3 piping stress analysis (MDOF analysis).



Pipe supports are in the flow chart as an added activity but for pipes designed using SDOF it is necessary to check pipe supports for imposed

pipe deformations in conjunction with applied load (see Section 6.3.4).



Equipment layout in CAD

Select design pressure vs criticality rating

Select scenarios Representative

of headline pressures

Specialist explosion pressures analysis

Zone module according to drag severity

Assume DLF =

RDLF = 1.0

Select event return periodvs criticality rating

Piping passes code Check with UF< 0.7

Layout and Piping Design

Implement lines into CAD

Develop line shoots

Process requirement

Process Design

Piping operating stress analysis eg Caesar II

Develop criticality ratingLevel 1 (high) & Level 2

Safety & explosion loadingResults of safety analysis QRA etc

Category 1 Pipe operatingPlus blast stress analysis

Eg Caesar II

No

Pipe supportDesign complete

yes

Select fire scenarios and consequence analysis

Develop barrier / spread philosophy

Select pipes for blast analysis

Does pipe pass code check for operating loads

yes

Generate drag loading by zone and direction X, Y, Z

Objects < 0.5m & all pipe


Category 1 pipe stress analysis

Structures

Deck & frame structure Design for explosion

Quasi-static Frame analysis(usually linear FEA)

SDOF analysis for criticality

Level 1 & 2 blast

Results of structural analysis

Non-critical supports(B team)

Changelayout

Design critical pipesupports (A team)

Support to category 2 and 3pipe stress analysis

Category 2 and 3pipe stress analysis

No

Optimiseglobal & local layout

No

Piping Design complete

Criticality 1

P & ID

Operatingparameters

Structural geometry

Drag / overpressureratio

Select criticality 1 and 2 pipes

Agree quasi-static design Pressure for category 1 type pipe

Stress analysisPeak drag pressureX, Y, Z directions

Relative movement of decks, pipe

supports etc

Criticality 2 & non-critical

Pipe support criticality rating

yes


Select design pressure vs criticality rating



Specialist explosion pressures analysis

Zone module according to drag severity

Assume DLF =

RDLF = 1.0

Select event return periodvs criticality rating

Piping passes code Check with UF< 0.7

Layout and Piping Design

Implement lines into CAD

Develop line shoots

Process requirement

Process Design

Piping operating stress analysis eg Caesar II

Develop criticality ratingLevel 1 (high) & Level 2

Safety & explosion loadingResults of safety analysis QRA etc

Category 1 Pipe operatingPlus blast stress analysis

Eg Caesar II

No

Pipe supportDesign complete

yes

Select fire scenarios and consequence analysis

Develop barrier / spread philosophy

Select pipes for blast analysis

Does pipe pass code check for operating loads

yes

Generate drag loading by zone and direction X, Y, Z




Structures

Deck & frame structure Design for explosion

Quasi-static Frame analysis(usually linear FEA)

SDOF analysis for criticality

Level 1 & 2 blast

Results of structural analysis

Non-critical supports(B team)

Changelayout

Design critical pipesupports (A team)

Support to category 2 and 3pipe stress analysis

Category 2 and 3pipe stress analysis

No

Optimiseglobal & local layout

No

Piping Design complete

Criticality 1

P & ID

Operatingparameters

Structural geometry

Drag / overpressureratio

Select criticality 1 and 2 pipes

Agree quasi-static design Pressure for category 1 type pipe

Stress analysisPeak drag pressureX, Y, Z directions

Relative movement of decks, pipe

supports etc

Criticality 2 & non-critical

Pipe support criticality rating

yes

Figure 6.1 Flowchart for design of criticality 1 piping



Pipes not passing category

1 pipe stress analysis


DLM analysis to develop loads for pipe-rack response anal;ysis

Select DLF and determine revised Quasi acceptabble-static design load

Does design pass SDOF analysis Engineering review of

acceptability of overload.

Layout and Piping Design Safety & explosion loading

Safety analysis QRA etc

Results of category1 operating &

blast stress analysis

Pipe support Design complete

yes

Check and amend pipe support beams


Structures

Quasi-static/SDOF analysis of major piperacks

Perform SDOF analysis & producecharts for drag DLF by zone

for use by piping group

Structural analysis

Local layout optimisation Possibilities ?

Design critical pipe supports for up-to date design loads

(support A team)

Category 3specialist pipe stress analysis

Explicit NLFEA

Not acceptable

No



Identify overloaded components, candidate failure modes & degree

of overload

Modal analysis for Natural period

No Is simplified quasi-static / SDOF

analysis appropriate

yes

yes

MDOF analysis ofpipe-rack structure

Does design pass MDOF analysis

Pipe rackDesign complete

acceptable

yes

Pipe rack Structural

modificationNo

No

Does design pass Category 3 analysis

yes

PipeDesign complete

No

Input support structure deformations as imposed

support movementCategory 2 pipe stress Analysis (eg Caesar II)


Pipe rack responseanalysis

Process Design

Local piping design

modification

Pipes not passing category

1 pipe stress analysis


DLM analysis to develop loads for pipe-rack response anal;ysis

Select DLF and determine revised Quasi acceptabble-static design load

Does design pass SDOF analysis Engineering review of

acceptability of overload.

Layout and Piping Design Safety & explosion loading

Safety analysis QRA etc

Results of category1 operating &

blast stress analysis

Pipe support Design complete

yes

Check and amend pipe support beams


Structures

Quasi-static/SDOF analysis of major piperacks

Perform SDOF analysis & producecharts for drag DLF by zone

for use by piping group

Structural analysis

Local layout optimisation Possibilities ?

Design critical pipe supports for up-to date design loads

(support A team)

Category 3specialist pipe stress analysis

Explicit NLFEA

Not acceptable

No



Identify overloaded components, candidate failure modes & degree

of overload

Modal analysis for Natural period

No Is simplified quasi-static / SDOF

analysis appropriate

yes

yes

MDOF analysis ofpipe-rack structure

Does design pass MDOF analysis

Pipe rackDesign complete

acceptable

yes

Pipe rack Structural

modificationNo

No

Does design pass Category 3 analysis

yes

PipeDesign complete

No

Input support structure deformations as imposed

support movementCategory 2 pipe stress Analysis (eg Caesar II)


Pipe rack responseanalysis

Process Design

Local piping design

modification

Figure 6.2 Flowchart for design of criticality 2 piping



6.3.1 Category 1 analysis: Basic ASME B31.3 code check procedure

Design method and applicable codes The basic analytical procedure for piping is a stress analysis to ASME B31.3 “Process piping systems”. This is a design code developed for piping systems within refineries and chemical plant but is used for offshore facilities too.

ASME B31.3 is a working stress code which gives allowable working stresses as a function of yield and ultimate stress, typically one third UTS or two thirds yield whichever is less. For bolting systems (flanges and bolts) a more onerous requirement of one-quarter UTS or two-thirds yield is given and this is partly to give assurance against leakage.

Explosion is often treated as an “occasional load” and combined with sustained loads (pressure and weight) but ignoring thermal stresses. For blast condition, a 33% overstress is usually allowed (see Example 1).

For bends and tees, stress intensity factors (SIF) are applied to simple or nominal elastic bending stresses based upon straight pipe. This takes into account the uneven stress distribution and ovalization effects which occur under pressure or external bending loads. For the allowable stress check, the nominal stress is multiplied by the SIF.

Piping stress analysis is a working-stress check, hence stresses have to be determined using an elastic finite element model of the pipe system, normally a stick or beam model, with nodes at potentially critical points. The most commonly used software is Caesar II and Example 1 is performed using this software. Forces are printed out and stress checks are made at each node. For pipe bends a node is normally placed in the middle of the bend, where the SIF is usually applied.

Nodes are also located at flanges and nozzles. For the treatment of vessels, it is necessary to specify fixed and sliding saddle locations to establish thermal stresses and the effects of piping loads on saddle support loads.

ASME B31.3 is conservative for bolt loads in flanges as leakage is the target failure mode. Maximum allowable bolt load is limited to 0.25

UTS (+33%). ASME DIV 1 (Nuclear design code) is sometimes applied for flanges and their bolting and, in the case of explosion, there is a Code level “C” which applies for seismic conditions which could be applied for explosion load. Example 1 includes a flange check on this basis.

Loading The explosion load can come from any one of 3 directions (two lateral X, Y and one vertical Z), with its intensity varying according to the dominant blast wind direction and the range of scenarios considered. The envelope of explosion wind loads in each direction may be obtained in a manner proportional to the level of criticality rating of the piping system being designed and the associated risk (see Section 3.3).

In most offshore situations the vertical loading is small due to the confining effect of the decks, but this may not be the case in open areas (e.g. FPSOs and onshore plant).

In Caesar II it is possible to input the loading in all three directions, and different load combinations may be used.

Windage area and drag coefficient The area over which the loading is to be applied is the pipe (or fitting or valve) projected area multiplied by an appropriate drag coefficient Cd. In general drag coefficients for cylinders may be less than 1.0 but given that form drag is not the only loading component for larger obstacles a conservative approach should be taken for obstacles (pipes and valve components) of diameter or width greater than 0.3m.

The Baker values (Table 3.4) have a level of conservatism so in the absence of better data these should be used for the larger items of pipes and valves, but not for objects over 1m (for which DLM method is more accurate).

In Caesar II the Cd is put in as “wind shape” factor, (see Example 1) and, where appropriate, should be augmented additionally by a factor = insulation OD / pipe OD.



Dynamic load factor Loading is applied as a quasi-static load, i.e. an equivalent static load. Normal practice in this form of analysis is to ignore dynamic load factor (DLF) i.e. set its value at 1.0 and rely on the inherent conservatism of the B31.3 code check criteria to cover for the fact the DLF is likely to be more than 1.0.

Rebound dynamic load factor The inbound response of the pipe is followed by a rebound response and this rebound case also needs to be checked. In open congested modules this rebound can occur at the same time as the reversal in drag flow direction and lead to higher loading in the rebound direction. This needs to be accounted for even in a quasistatic analysis by having a rebound case. For most pipe situations rebound resistance is the same as resistance in the primary direction. For pipe supports it is recommended to make design rebound loads equal to half the inbound load unless SDOF analysis demonstrates the suitability of a different value for the rebound direction.

Screening for pipes which may need further analysis The design process relies on assumptions that may be conservative and others that may be unconservative; and it is often assumed that there is probably no systematic bias in either direction. However, for safety reasons it is recommended to identify piping components (using Category 1 pipe stress analysis as shown in Figure 6.1) that have utilization factors over 70%. These components will then be analysed using Category 2 piping stress analysis (Figure 6.2) to determine whether the dynamic load factors applicable to them are so much more than 1.0 that further analysis or design (Category 3 pipe stress analysis) is required to justify them.

This approach is also essential for the pipe support design, which, being geared to the actual support load, will often systematically come into the 70% plus range and often has lower safety factors (see Section 6.3.4).

6.3.2 Category 2 analysis: Conventional piping stress analysis enhanced with SDOF analysis

SDOF analysis and hence Category 2 type analysis is only applicable to simple structural

systems such as straight pipe with particular support conditions. It is also applicable to more complex structural systems that can be broken down into simple subsystems that can safely be assessed in isolation. In other cases, Category 3 MDOF analysis is required.

The objective of the SDOF analysis is to quantify the dynamic load factors to apply to quasi-static explosion loads in a more refined re-run of a Category 1 code check analysis.

The primary screening tool is a Caesar II analysis or equivalent as described in Section 6.3.1 above, using Category 1 pipe stress analysis. Those components having for instance more than 70% utilization factor (UF) should then be systematically, and individually reviewed to see whether the simplifications made in the quasi-static analysis could outweigh the inherent conservatisms of the ASME B31.3 code check. The same assessment process is applied to components which have utilization factors above 100%: it can often be shown that overload factors can be accepted in certain circumstances.

The following information is required:

1. An estimate of the dynamic load factors (DLF) and rebound (RDLF) for explosion scenarios typical of the return period explosion loading that has been set for the criticality rating of components being assessed. As DLF varies with period of vibration TN and allowable plastic deformation the DLF and RDLF values are presented as graphs of DLF vs. TN.

2. A statement of the failure mode(s) which the Caesar II has identified as critical for each overloaded component of the pipe system (for which UF>70%).

3. An assessment of how the component will behave in the identified overstress situation. For instance if it is an SIF problem on a bend in ductile material and coexisting operating pressure is well below allowable (as a load case on its own) the degree of allowable overstress could sometimes be very large and aspects such as strain rate enhancement of the material and yielding may be allowed.

4. A guide stating which failure modes will happen following ductile bending of the



pipe - in which case the lower DLF values from 1) above will apply.

Estimating DLF and zoning according to severity of drag pressure The method adopted will depend upon the geometrical configuration of the piping and how the explosion loading is defined. In general the explosion loading must be presented as a pressure time history to each applicable loading direction (e.g. x, y and z).

Where explosion loads are defined probabilistically, this means selecting one or more explosion scenarios representative of the design accidental event. This part of the work is normally one of the functions of the design accidental loads specification.

Each of the representative pressure time histories (usually drag-pressure time histories) is applied to an SDOF model several times with the natural period of vibration of the model altered for each simulation. The dynamic load

factor DLF is equal to the peak resistance mobilised in the spring of the SDOF model divided by the peak value of the load applied to the model. There are a number of sequential peaks of the response and the ones to use are the maximum ‘inbound’ response and the maximum rebound response. The values will be different according to the natural frequency of vibration of the SDOF model.

In practice, pipe runs and systems have a variety of different natural periods hence the DLFs need to be presented as a function of natural period of vibration and allowable ductility ratio. Figure 6.3 shows an example of this for one blast impulse shape in one zone [6], where the expression Mu is the ductility ratio μ. A family of such curves is required to cover a range of scenarios in any one zone of a platform. As X, Y & Z direction drags need to be treated separately, the number of curves is potentially multiplied by 3, but the Z direction drag is usually small enough to be ignored.



Drag pressure-time history

-3.5-3.0-2.5-2.0-1.5-1.0-0.50.00.51.0

0 200 400 600

time (ms)

P (k

N/m

2)

(a): Drag pressure time history

DLF vs Natural period (Scenario 1)Elastic and elasto-plastic response

-2.500

-2.000

-1.500

-1.000

-0.500

0.000

0.500

1.000

1.500

2.000

2.500

0 50 100 150 200 250 300 350

TN (ms)

DLF

DLF+ 'elastic

DLF- elastic

DLF+ 'Mu = 5

DLF- Mu = 5

(b) DLF as a function of natural period

Figure 6.3 DLF and associated blast impulse [6]

Figure 6.4 DLF for triangular impulses [58]



In many situations in zones where drag loading is highest there is a strongly dominant drag direction and drag pressures transverse to this direction are small; the transverse pressures can be estimated as a proportion of dominant drag and safely assumed to have the same DLF.

It has been observed that the impulse shapes for outward drag near the boundaries of semi enclosed modules are largely triangular enabling the use of Biggs curves or, for isosceles shaped impulses, Figure 6.4 where T is the natural period of the structure. FABIG Technical Note 4 Appendix A gives rebound DLFs for some common impulse shapes.

For the internal zones of modules where impulses are weaker but more complex impulse shapes apply, one may be able to avoid analysis, simply take a blanket DLF of 2 and still have a reasonable design quasi-static load. Whether this is practical or not depends upon the severity of the design blast load (headline pressure) and corresponding design drag pressure.

In FPSOs and onshore plant the drag impulses are relatively high (as a proportion of headline design pressure) and at the same time have complex shapes. In this case Category 2 or 3 analysis is required and it should be realised that rebound DLFs can be bigger than inbound DLF, ie the pipes could break due to “forces” in the opposite direction to drag flow.

The designer needs to define the mode shape that most closely matches the deformed shape of the pipe under the (critical) blast loading and use the associated natural period obtained from simplified formula in interpreting the DLF curves. Caesar II and equivalent codes may be used for more complex geometries (e.g. IGN, TN4, Figure 6.2).

Allowable overstress and plastic deformation of pipes and bends For overstressed line pipe or bends, one can expect the pipe to buckle or bend plastically before rupturing and this means that a degree of overstress can be tolerated for the explosion condition. Exceptions are brittle materials such as GRP and situations where buckles can form near girth welds, such as weld-neck flanges, because the bend line at the edge of the buckle can form at the weld leading to premature failure (loss of containment).

Buckling strength can be determined by reference to specialist literature or by suitable NLFEA. Buckling reduces resistance and when this happens while the loading is still applied, complete failure of the component follows rapidly. Elasto-plastic response is only allowable when buckling or other resistance reducing effects do not occur.

Deciding on ductile bending (rotation) capacity If a pipe yields and deforms plastically (in bending) before rupturing or otherwise failing, more energy is absorbed and dynamic load factors will be lower. To evaluate elasto-plastic DLFs, the SDOF model can be enhanced by allowing it to deform plastically. The measure of plastic deformation that is used in the analysis is the ductility ratio μ which is the maximum deflection of the pipe span in bending, relative to its supports divided by the deflection of the same reference point at the elastic deflection limit (the limit of load deflection proportionality).

For convenience, DLF vs. natural period curves are prepared as a family corresponding to various ductility ratios (see Figure 6.1b).

It is then necessary to decided on the allowable ductility ratios for a pipe span. If the pipe has flanged connections, tees or branches that are located in highly stressed regions, then the ductility ratio should be limited to 1.0 and hence the analysis should be limited to the elastic domain. For straight pipes between strong supports, considerable levels of ductility ratios may be allowed.

In ENV 1993-1.1 Table 5.3.1e, plastic bending of pipes can occur without buckling if d/t<50ε2 where ε = (235/fy)0.5 where fy is yield stress or RP02. This is applicable for ductility ratios up to 1.5. For higher ductility ratios a check needs to be made to Reference [59].

Where d/t>90 ε 2 the full elastic bending strength of the pipe cannot be mobilised without premature buckling and ductility ratios must be restricted to 1.0, with strength determined with due account for buckling. Strictly speaking one should consider the pipe without corrosion allowance, unless it can be shown that corrosion would be sufficiently localised that buckling strength would not be affected.



The above buckling limits do not take account of internal pressure, which reduces buckling risk and increases allowable ductility ratio. In flare and firewater lines, internal pressure at the time of explosion will be small hence buckling risk should be assessed on the basis of unpressurised pipe.

High internal pressure might, if it is contributing heavily to hoop or longitudinal stress, affect bending strength at yield and possibly limit ductile bending (rotation) capacity. Theoretically the effect should be small (there is only a small increase in Von Mises stress) but tests and analysis would be needed to confirm this.

Where plastic bending deformation is allowable, e.g. straight pipe in racks, a lower level of DLF may be used consistent with the allowable ductility ratio. However, limits on span shortening need to be checked.

The plastic moment capacity of metal pipes (without internal pressure) is fy

* x 1.333 (R3 - r3). Where fy

* is the yield stress enhanced for strain rate (typically by a factor 1.1 to 1.2) and R and r are the outer and inner pipe diameters respectively. Where it is found in the analysis that moments are exceeding this value (overstress situation), plastic bending (rotation) is occurring and load redistribution will take place affecting support loading. This should be taken into account where such effects could lead to increase in support load.

Special component studies Where tees are located in critical lines (including pipes with weld-o-lets) weak points are introduced. Orientating the weak points to neutral axis level (for dominant drag loading) is sometimes a possibility. As regards tees one can specify hot forged tees of nominally thicker material and extra thickness in high stress areas. Such actions can increase allowable stresses and, if necessary with the support of FE analysis, can lead to tee sections having strengths equal to the line pipe they are connected to.

Some piping problems boil down to finding out the ductile bending capacity (moment rotation curve) for a particular nozzle or a large Tee in a riser (e.g. before a pig receiver) or a simple pipe bend. These are analysed by a shell or solid element component model (NLFEA) and it is necessary to check local strains so as to identify local failure risk. Implicit NLFEA cannot handle plasticity and buckling. For such tasks one needs

to use an explicit type such as ABACUS Explicit or LSDYNA.

6.3.3 Category 3 analysis: MDOF analysis method

Complex structures and pipe systems requiring MDOF analysis In general the vast majority of pipe systems on an installation will have been confirmed and accepted on the basis of the simpler methods already described. Some systems will require further analysis due to the combination of complex geometry and high blast loading. In such cases analysis to ASME B31.3 is not suitable and it is recommended instead to use Multi Degree of Freedom (MDOF) analysis.

In this context, complex structures are:

• Large pipe racks, i.e. the whole assembly as opposed to individual pipes within them;

• Pipe racks where plastic analysis is required in order to demonstrate ductile behaviour and lowered DLFs;

• Pipe runs with particularly difficult geometries, changes of size, branches etc, which makes it difficult to justify using SDOF analysis;

• Pipe runs with important weak points such as pipes from nozzles of gas compressors or vessels where one requires an accurate value and assessment of estimated nozzle forces (e.g. for the equipment designer);

• Where ductile bending capacity has to be relied upon. This must be confirmed, usually using an acceptance criteria of maximum allowable local strain; and

• Structures where deck movement imposes forces on the pipe and pipe forces that have to be estimated in the time-domain.

Pipe racks For pipe rack structures, the whole structure and pipes can be modelled as a frame model. NLFEA models such as ABAQUS, LSDYNA and USFOS, amongst others, may be used. Programs for concrete pipe-racks exist too, usually having their origins in seismic design of framed buildings.

Loading on pipe rack structures is discussed in Section 6.3.4 and 6.3.8. Large complex pipe racks will have to be analysed for specific explosion scenarios as impulse shape, load



distribution and pressure variation along and across the structure will be an important consideration. Pipe rack structures may be analysed, at least initially, using stick models.

Pipes within pipe racks are generally of a simple form of construction and do not themselves require MDOF analysis. However, they do contribute to the pipe rack model, especially the loading, and therefore they do need to be modelled. In addition, pipe spools going out laterally to adjacent modules or plant plots will need to have the relative displacements of their ends properly determined so that they can be checked on an imposed strain basis.

The differential pressure across the pipe rack structure will probably have to be translated into a “wind” load, which, when shared between the members, gives the same overall force difference across the structure as the loading from the DLM analysis. The impulse shape will be the same as that for the overall differential pressure found in the DLM analysis. The CD factors have to be included and adjusted according to member shape, pipe grouping effects and shielding effects.

MDOF analysis methodology The initial model will be a non-linear FE model, where plasticity and strain is calculated. The modelling will not initially include details which enable local buckling or fracture of weak connections to be assessed. The first step is to carry out an elastic analysis, check all components and list those that do not meet code requirements, due to overstress or where plasticity is occurring. The analysis would then be extended into the elasto-plastic range with a code check against, for instance, an earthquake design code such as Eurocode 8.

In some codes it is possible to specify P-Δ curves for members and moment-rotation curves for joints (or short members near joints). In most models, the ductile capacity limits and non-linearity in the moment rotation curves is user-specified and this means resorting either to component models to check the suitability of the curves and ductility (failure) limits or alternatively by references to literature. This is a specialist and difficult area.

Component models usually have to be shell element models run in an explicit code. Sometimes it can be better to do the whole frame

analysis in an explicit code, otherwise it might not be possible to complete an analysis problem satisfactorily. The more sophisticated models (e.g. ABAQUS and LSDYNA) allow the component models to be integrated into the frame model, thereby removing most of the human intervention element in the analysis. This is a specialist area, particularly when it comes to interpreting results and assessing whether suitable component and local performance criteria are met.

Pipes normally have a much shorter vibration period than the rack structures they are supported in and consequently much higher DLFs for the type of loading under consideration. The pipes should be checked separately using their own DLFs and the supports and support beams designed for these loads (see Section 6.3.8 on pipe racks). In many cases it is adequate to assess these pipes entirely with the quasistatic approach in Ceasar II but it would be necessary to implement imposed support movements, consistent with the results of the MDOF analysis of the whole pipe rack frame.

Flanges and bolts Flanges may have strengths in bending which are less than the pipe in which they are located. In ductile piping systems this will be important where such flanges are located near points of maximum bending moments.

Bending resistance assessments need to be made both with and without corrosion allowance in the pipe ( for the case where the corrosion allowances are not subtracted the applied bending moments are higher). The resistance of the bolted flange should be at least 20% more than the moment of resistance of the pipe (see Interim Guidance Notes, Section 3.5.4 [2]).

Where stresses due to internal pressure are high bending resistance might be reduced, either in the pipe or the bolted flange and by different proportions. To determine the reduction one may carry out a Von-Mises stress type integration around the pipe where the bending stress variation around the pipe is determined so that Von-Mises yield criterion is reached at all locations around the pipe (an upper bound resistance is being sought).

The flange pressure and bending resistance assessment to ASME B31.3 will always show



the bolts to be overloaded, because the stress limit is 0.25 UTS, based upon simple elastic approach to determining bolt force. ASME DIV 1 gives less onerous criteria but these may not be met either.

The physical problem is that, in relation to bolt stiffness the flanges are relatively flexible and uneven load distribution between bolts will occur. The type of flange face and gasket also has an influence and prying effects occurring in the bolts, especially with a flat-face flange. To avoid such situations, it is recommended to use raised face flanges. It is not clear how hub connectors can be checked to ASME B31.3 as the calculation of bolt force is not simple.

Finite element analysis of flanges is possible but complex to carry out, as modelling bolts is not straightforward in FEA of bolted assemblies.

Bolts are of ductile material, the most common ASTM A193 B7 bolts, which are of chrome-molybdenum steel with minimum elongation at break of 16% and 50% reduction of area. UTS is 690 N/mm2. The bolts are threaded throughout all (or at least part of) their tensioned length and therefore have a series of stress raisers (each thread root). This limits their ductile elongation capacity and the ability to respond to prying effects by straining.

At this time, no simple methodology can be confidently stipulated for assessing the bending strength of flanges other than the criteria of ASME B31.3 or ASME DIV1.

A solution often used in structural applications is to have only the ends of the stud bolts threaded and the unthreaded part turned or machine rolled down to a diameter less than the thread root diameter. This ensures that longitudinal straining of the bolt will cause elongation in the shank zone where there are no stress raisers. This allows a simplified approach to calculating load distribution between flange bolts, without the need to calculate imposed strain and a more optimal bending resistance of the bolted flanged joint.

6.3.4 Pipe supports Project organization Pipe supports have both piping elements (treated in the piping design code) and structural elements (treated in a structural design code). The vast majority of supports are very simple

and designable at designer-draughtsman level but some are highly critical and need very careful attention. Some of the more difficult design issues are discussed below. This needs to be recognised in the organisation of the pipe support design work and the screening process that decides which Category a particular pipe support fits into.

Given these factors and the traditionally low level of design effort (and capability) associated with pipe support design it is recommended to screen pipe supports in accordance with piping criticality and split the pipe support team into two parts. A first team with the knowledge and capability to understand the structural implications of all the necessary factors and the second level group who can remain as a rapid design turnaround team for the simple support and subsidiary to the piping stress group. It is proposed that the first team should probably report to the structural discipline and liaise with the piping stress group rather than the reverse.

There is another issue that is actually very much in line with the objectives of weight and cost efficiency. This is that unnecessary tertiary steel causes added turbulence in an explosion and increases explosion overpressures. Efficient and compact design of pipe supports is therefore to be an aim, and this is probably best satisfied by putting effort into producing a good standard set of project pipe support details for the pipe support designers to adapt their designs from.

Support design methodology Blast loads are directional but there is in all cases a primary load in the direction of the drag load and a rebound load in the opposite direction: for each blast load direction there is therefore a minimum of two load cases one primary and one rebound.

Loads for pipe support design are generally produced in Caesar II as X Y and Z direction loads (see Example 1, output pages 6 & 7). Pipe support design is split into two parts: the part that is designed with the pipe (using Caesar II) and the part that is designed as part of the structure, using structural design codes.

Both the pipe supports and the structures they are connected to need to be verified for the loads applied by the piping.



The conservatism inherent in the allowable stress approach in ASME B31.3 does not exist in the structural codes. It is important that the structural designer recognises this when interpreting the design support loads. The protection against ignoring DLF (setting it to unity) is there for the parts of the piping support system that are designed conservatively to ASME B31.3 but not for the remainder. Therefore, support loads might sometimes be unconservative.

In explosion resistant design (see Interim Guidance Notes Section 3.5.4) it is recommended to design critical supports for loads acting upon them from the structural components they support and factored up by 1.2. This is to give protection so that failure in one critical support does not change the load distribution in the whole structural system thereby leading to a cascade of pipe support failures.

In the event that engineering judgement is used to justify a higher allowable blast loading for the piping component design, Caesar II analysis must be re-run for using the enhanced loading so as to generate new design support loads and document the pipe support design using them. In other words the pipe support design must be kept up with changes in the pipe stress analysis and this must be carefully documented, especially for the critical first-team pipe supports.

Very often, a pipe itself is dominated by drag from one direction: drag flow across the pipe. It is often possible to treat the X Y and Z direction drag load combinations separately and make sure each individual load combination gives satisfactory stresses. This is what has been done in Example 1 but the validity in this particular instance has not actually been explored. In some instances it may be necessary to consider different load combinations.

The support however has a more complex loading: and for pipe supports it might be unsafe not to create the “grand” combination sum of all three (X Y and Z blast combinations) with different multiplication factors for each.

An additional problem is that under blast loading the pipe may rotate about X, Y or Z axes, creating an uneven interface loading. In guided supports this may necessitate adopting ductile

form of construction: i.e. by avoiding local buckling (minimum plate d/t ratios) and sizing welds so that enforced displacement of plates meeting at welded joints does not break the welds. Prying in bolted connections (see ENV 1993-1.1 cl. 6.5.9) can also become important. In normal pipe design movements are kept small and these considerations can usually be ignored but in critical pipes such imposed movements should be reviewed.

In fixed supports where large pipes are restrained the pipes may move relative to the support structures in a way that forces a distortion of the pipe support. Pipe loads (from Caesar II) may be insufficient to cause this problem and the pipe support can be designed statically but if the support is moved relative to the pipe (e.g. due to deck movement due to blast) these forcing effects can be significant because of the rigidity of the pipe: a hybrid support may be needed, i.e. one which can withstand the pipe loads but also can be twisted or otherwise deformed when the deck moves without breaking the support or overloading the pipe. This sort of load is much more difficult to account for in Caesar II / ASME B31.3.

Where piping is reanalysed by MDOF methods, pipe support loads need to be checked within the MDOF model and will necessarily be scenario based. Sometimes a large number of scenarios with equivalent return probability will have to be considered and an envelope of support loads created. For pipe racks the racks themselves will respond one way to the blast impulse (at a natural period corresponding to excited vibration mode of the rack assembly) but the pipes within it will vibrate at their own natural frequency. The individual pipes will invariably have a significantly shorter natural period of vibration than that of the pipe rack and in consequence their DLF will not be much affected by overall movement of the pipe rack. Normally, the pipe’s DLF should be used for the pipe design and the rack’s DLF for the rack design.

One case where this approach may be problematic is for vertical fronted shock waves where pipe support reactions may need to be calculated according to Biggs. The shock loading impulse can be determined according to Kinney and Graham [52].



6.3.5 Enhanced design In all cases the approach must be one of practicality. Where a large number of pipes have to be designed to withstand explosion, reducing headline design pressures and drag forces across pipes by adopting inherent safe design practices will reduce the problem down to a scale that can be handled by simple quasi-static techniques (Category 1 analysis). This broadens the number of potentially critical items that can be checked and confirmed.

It is frequently the case that a project will have a few pipes with considerable design problems and these will usually be identified from results of the quasistatic approach (Category 1 analysis). In the case of few such pipes, it becomes practical to target resources on them to improve their design.

For enhanced resistance there are four parallel routes to be followed:

1. Good overall layout optimisation so as to reduce the 10-4 headline pressure;

2. Good local layout design, locating critical items in “lower drag” areas and reducing congestion due to tertiary structure (avoiding excessive or wasteful pipe support provision);

3. Building in strength and ductility; and

4. Optimum pipe rack design.

6.3.6 Strength, ductility and robustness In Sections 4.1 to 4.3 it has been shown that judicious layout design and pipe routing reduces the required quasi-static design load for piping, for a given headline design field pressure.

Building in strength will increase the quasi-static load capacity of piping systems.

Building in ductility and judicious pipe routing will enhance mostly its dynamic blast capacity. By allowing a higher ductility ratio to be used, this will allow a reduction in design quasi-static design load for a given design headline field pressure.

The amount, weight and cost of pipe support structures installed are a function of the quasi-static load capacity that has been built in. This means that astuteness in pipe routing and building in ductility are potentially very cost-

effective measures. It should be recognised that reducing tertiary structure reduces turbulence and headline explosion pressure.

Robustness is about trying to achieve the goal of enhanced blast resistance by general practical measures rather than quantified means. Given the complexity of the problem (i.e. understanding drag, dynamic response and material behaviour) in some instances one cannot know instinctively what would be an improvement and what would not be. Consequently the qualitative activity of building in robustness cannot take place without the support of quantitative analysis or a written well-researched guide on how to achieve robustness.

6.3.7 Measures for increasing blast resistance

Increasing quasi-static resistance may be achieved using one or more of the following options:

• increasing pipe wall thickness, • increasing flange rating class, • avoidance of branch pipes and tappings in

parts of piping systems where bending moments are a maximum.

It is also about reducing support spacing but this can be in conflict with the desire to accommodate thermal movements, reduce fatigue stress and noise and accommodate deck movement in blast: it is sometimes not an option.

As regards accommodating deck movement in fixed offshore structures, reducing support spacing would have to go hand-in-hand with reducing deck movement due to blast, e.g. strengthening or installing floor to ceiling ties.

In the case of small piping it can be a good idea to stipulate minimum pipe sizes in racks so that they can inherently span the standardised support spacing; typically a lower limit of 2” (68mm OD) is practical but possibly more if the pipe is enlarged by insulation.

6.3.8 Measures for increasing ductility The dynamic load factors, DLF and RDLF, which are used to define the loading in the positive and the rebound phases, can be reduced by a factor of two or more if ductile construction



is used. This is particularly true for items with long periods of vibration. Increasing ductility may be achieved using one or more of the following options:

Material behaviour

• Ductile capacity is very dependent on the shape of the material stress strain curve in the strain hardening region;

• Piping response is more likely to be ductile if the material is ductile with yield to ultimate less than 0.7 or 0.8;

• Composites like GRP have a yield to ultimate tensile strength of 1.0 and therefore load reduction due to ductility cannot be taken into account;

• Using materials with ductile behaviour at low temperatures for those cases where low temperatures can occur under upset process conditions: especially blowdown and in or near low temperature jet releases; and

• Ductility of material does not always lead to a ductile structural behaviour because the failure mode can be brittle, e.g. buckling or connection weakness.

Structural behaviour

• Response and resistance is dependent on whether piping component can deform in a ductile way;

• Test data for pipes and fittings subject to ductile bending while under pressure does not appear to exist;

• Piping without fittings will often have an elasto-plastic response; however, the assumed ductility of the pipe should be checked against the fracture strength of the material;

• If plastic hinges form in components which are inherently ductile: A ductile bending behaviour can be assured for the line as a whole if the simultaneous bending moments in other components of the line that are weaker or are non-ductile are less than the static capacity of such components;

• If plastic hinges form in components which are inherently ductile and the bending moments in components that are weaker or are nor ductile are within the static capacity of the components then a ductile bending behaviour can be assured for the line as a whole;

• Locating flanged connections, tees and tappings away from points of maximum bending moment or locating them at the neutral plane;

• Increasing flange rating to make them stronger in bending than the pipe and assessing them with pipe wall increased by corrosion allowance;

• Having a pipe wall thickness so that the pipe d/t range falls into the ductile range (for guidance on tubes without internal pressure see FABIG Technical Note 4);

• For bends in high moment regions, having pipe bends with documented plastic bending capacity; and

• Careful assessment of the likely effects of corrosion to produce defects/pits/grooves transverse or longitudinal to pipe.

Bolts and welds

• Welds and bolts are loaded indirectly as they transfer the forces generated by an explosion from one member to another. The significant characteristics of these forces are that they occur rapidly and are large in magnitude;

• Welds normally have mechanical properties similar to or better than the steelwork to which they are attached. However, there is a greater variation in properties than for rolled steel and consequently “bad” zones can exist. These are particularly prevalent in the parent metal immediately adjacent to the weld (the heat affected zone). The result can be that small defects lead to brittle failure under rapid loading. In general, these problems are completely avoided by suitable detailing and correct welding procedures; and

• Provided bolts are sized to take the explosion-related load, there should be no problem with regard to premature failure. Where loads exceed the design value, bolts will either fail in brittle fracture or by plastic deformation. The higher the grade of bolt used, the more likelihood there is of brittle fracture. Bolts used offshore are usually Grade 4.6 or Grade 8.8 which should not be prone to brittle failure.

Flare systems and blowdown In regard to flare systems, flare lines are relatively large, mostly welded and consequently



quite strong, however being normally class 150lb, the flange connections are not necessarily as strong as the pipe. As the flanges have a degree of flexibility and the bolts are short they can be unevenly loaded due to bending loads in the pipe, which can lead to premature bolt failure. A potential solution in critical cases is to raise flange class from 150lb to 300lb.

Also the flare lines are not normally highly pressurised and are not likely to be so at the time of a blast. Given a diameter to thickness ratios of possibly 100 they will be prone to buckling, especially the bends. Plastic moment capacity will then not be available because the pipes will buckle under bending overload, weakening them. It has been known to have buckles forming next to weld neck flanges and leading to fracture at the weld so this form of failure is not without leakage risk. It should be noted that in many instances a moderate amount of buckling would not cause a leak in the pipe wall.

Another issue to be considered is low temperatures during blowdown (if it is activated before an explosion). This can reduce ductility leading to brittle fracture risk.

Many other line types and especially high pressure lines have low d/t ratios and will bend plastically before failure, especially those in ductile materials such as stainless steel or carbon steel A333 grade. With carbon steel sudden blow-down might be able to drop the carbon steel material temperature into the brittle range, leading to fracture and full bore release. This would be a risk if blowdown is initiated before an explosion or the pipe was affected by the jet-release leak.

Manifold and wellstream flowlines Manifold and wellstream flow lines tend to be high pressure, thick-walled and very strong and ductile, but they are difficult to route and routing sometimes has to be decided during operation due to changed sequence of well-slot allocation. These spools will often be connected with hub connectors and not all of these have bending capacity equivalent to the pipe, especially where wall thickness includes additional erosion/corrosion allowance: Again there is the difficulty of locating anything but the most rudimentary of pipe supports in wellheads areas (traditionally hangers and a few guides).

Given the high explosion turbulence (and drag) in such areas and the fatigue/thermal and wellhead movement effects that have to be allowed in design, it is preferred that an inherently ductile form of construction be adopted. The hub or compact flange connectors should be selected so as to be as strong in bending as the pipe that connects to them unless located away from points of maximum bending moment (in blast).

The wellhead nozzles need to be at least as strong as the pipe too, which is not likely to be a problem for the supplier given how strongly made the X-mas trees usually are.

Ductility issues and pipe racks In all cases the form of construction of the rack will be a critical aspect in terms of resistance and a ductile form is definitely recommended: with large pipe racks natural periods of vibration in sway may be as long as 0.5 seconds or more and if the construction form is ductile the dynamic load factor can be 0.4 or less (see Figure 6.3). In its most practical sense the Dynamic Load Factor (DLF) is the factor by which the peak dynamic drag force is multiplied to establish the design quasistatic load.

Figure 6.4 shows how DLF varies with td/T, for isosceles triangular impulses. td is the positive phase impulse duration divided by the period of vibration of the structure - for the mode that most closely resembles the deformed shape of the structure under quasi-static explosion loading. The ductility ratio is the maximum deflection (for the same mode shape) divided by the deflection of the structure at yield. The allowable limit on this is set by the ductile capacity of the members and joints that make up the pipe rack support frame and reference to an earthquake design code like Eurocode 8 is recommended for this. TN4 Section 7.2 gives simple guidance on ductility ratio limits for H section beams.

Usually one finds that drag impulses, though intense in these locations are of short duration. In response terms, structures with long periods of vibration and ductile (plastic) bending capacity have high tolerance for short duration impulse. The nature of the response of a relatively heavy pipe rack, constructed on the Vierendeel truss principal using class 1 beams and columns with ductile connections, is ductile



hence low DLFs and quasistatic loads can be achieved.

It is worth noting that unless structures and their joints are specifically configured for ductile response (Figure 6.5); they are not likely to have any significant ductile capacity, which means they could easily collapse if the elastic deflection limit is exceeded. The joints are usually the weakest link.

Figure 6.5 Pipe racks with ductile bending capacity for sideways drag loads

For the foundations of land-based plant, the base connections need to be moment resisting which can mean the use of ground anchors to keep footing sizes down, especially where uplift occurs in conjunction with lateral bending.

The pipe supports themselves will still see the un-attenuated design load, consistent with their higher frequency of vibration, so they and the beams to which they are connected should be dimensioned using higher DLF values eg 1.0.

Blockage In pipe racks a horizontal array of pipes can be dense: blockage ratio for vertical drag up to 0.8. In this case they will act like a deck and the pressures will be consistent with this: in such cases pipes need to be tied down for such forces, perhaps with ropes down and around the support beams.

On onshore plant pipes may be augmented in diameter by insulation, for heat conservation and noise reasons. This makes small lines quite big from the loading and blockage point of view.

6.3.9 Additional factors affecting the selection of method of dynamic analysis

The design flowcharts show that either SDOF or MDOF methods may be carried out for Category 2 pipe stress analysis. For the analysis of pipe racks the design flowcharts show that the analyst must chose the appropriate method of analysis.

The main problem with the SDOF method is that it is only feasible to apply to structures which can be characterised by a single stiffness and the loading by a single time varying curve. With overpressure loading, drag loading and forces due to relative movements of the supports, there are always at least two separate load quantities acting in different directions and with different time variation. In this case, errors are likely to arise when using SDOF methods and the procedure will become cumbersome and more expensive than MDOF analysis.

Selecting an appropriate method (SDOF or MDOF) for the analysis of the response of piping systems subjected to explosions, is dependent on several factors including:

• Loading regime (relative period of loading to period of piping);

• Coupling of piping/topside response (relative period of piping to period of topside structure); and

• Validity of superposition principles.

Loading regime The Interim Guidance Notes [2] defined different loading regimes based on the duration of the loading as compared to the natural period of the structures. The boundaries of these loading regimes were then revised by the Norsok Standard N-004 Design of Steel Structures – Annex A [60], and are reproduced in Table 6.1 below:

all are moment fixity joints with ductile rotation capacity



Table 6.1 Effect of loading Regime on treatment of overpressure component of loading [2 and 60]

Loading regime Parameter

Impulsive (short) τ/T < 0.4

Dynamic (intermediate) 0.4 < τ/T < 3.0

Quasi-static (long) τ/T > 3.0

Peak Value Preserving exact peak not critical

Increase or decrease in peak will result in a similar trend in the response

Duration Preserving exact duration not critical

Slight changes in duration may affect response

Duration more important as response becomes plastic

Impulse Accurate representation of impulse important. Negative impulse may be important

Accurate representation of total and rise times important

Accurate representation of impulse not important

Rise Time Preserving rise time not critical

Preserving rise time very important.

Idealised pressure time

Analysis method Energy methods SDOF / MDOF Static / energy

methods Notes: 1. τ is duration of loading 2. T is natural period of piping system 3. tr is rise time of loading

Coupling of piping / topside response The nuclear code ASCE 4-98 [61] provides clear guidelines as to the criteria for selection of coupled/uncoupled analyses for items of equipment. ASCE 4-98 [61] recommends that coupled analysis is not required if the equipment (or secondary system) satisfies the following requirements:

• Total mass of component is 1% or less of supporting primary structure. If components are identical and located together, their masses shall be lumped together;

• Stiffness of component supported at two or more points does not restrict movement of primary system; and

• Static constraints do not cause significant redistribution of load in primary structure.

For tanks and vessels with single deck attachment (i.e. connected at one deck level only and with no significant separation between the support points so that the acceleration at the various points can be assumed to be the same),

the selection of coupled analysis or uncoupled analysis is based on the frequency ratio and the modal mass ratio. From the numerical values of the frequency ratio and modal mass ratio, the selection of the type of analysis can be carried out based on Figure 6.6[61].

Validity of superposition principles If a SDOF method is to be used then the response due to drag, overpressure and differential will have to be determined separately and then summed using superposition principles. Even for small diameter piping where overpressure can be ignored, there may be two components of loading (drag combined with deck accelerations) in different directions and with different time variations. The applicability of SDOF methods in such cases becomes questionable. The situation becomes worse if one is to take plastic deformations into account; however, FABIG Technical Note 7 [62] provides a SDOF method which takes into account catenary effects. OTO 1999 046 [6] concludes that, in practice, the drag loading is

P

t

τ

tr

P

t

τ



largely confined to one direction and the problem in applying the SDOF method is reduced to how to combine a drag load in one

direction with differential deck acceleration effects.

Figure 6.6 Selection of uncoupled versus coupled piping – topside analysis [61]

In cases where the drag loading is confined to one direction, and where the overpressure loading is negligible, the problem in applying SDOF method is reduced to how to combine a drag load in one direction with deck acceleration effects. This becomes more of an issue when non-linear effects have to be accounted for, to take advantage of the ductility of the system.

In all cases the forces caused within the piping item and its supports are dependent upon:

• the mass and stiffness of the piping system; • the stiffness of the support structure; and • the ductile deformation capacity of the

piping system and its supports.

6.3.10 Structural treatment by SDOF The objective of the SDOF method is to translate a dynamic pressure or load into an equivalent static load for the piping and equipment design. The equivalent static loads are the peak dynamic loads found in a time domain dynamic response simulation of the structure under the applied dynamic loading. It is important to evaluate the maximum positive

and negative deflections and forces. The time domain simulation requires the preparation of two separate idealised models of the piping system item and support structure [6]:

1. Apply the peak explosion load to the model multiplied by a suitable dynamic load factor (DLF);

2. The negative or rebound response will be a further load case where the peak explosion load is multiplied by the rebound dynamic load factor (RDLF), which has the opposite sign of the DLF;

3. Treat the loading in the X-, Y- and Z-directions as separate load cases;

4. Finding DLF and RDLF requires the modelling of piping and its surrounding support structure;

5. The SDOF model is an idealisation of the piping and its support structure, where the spring has a stiffness K such that the mass Me has the same deflection as the mass centroid of the piping item when subject to the same load. Mass Me is the mass of the equipment multiplied by a transformation



factor whose value is such that the period of vibration of Model 2 is the same as that for Model 1 for the appropriate vibration mode shape (see Figure 6.7); and

6. The X- and Y- direction forces are drag force components, the Z-force acting is the differential pressure across the deck. The mode of vibration will be different for each direction hence the natural period of the SDOF Model 2, the spring stiffness and

mass Me will be different for each of the directions. The force time histories will have different shapes and maxima. It is necessary to consider X-, Y- and Z-directions separately and obtain DLF and RDLF for each direction and then sum the worse values of DLF and RDLF for application in the SDOF model.

(a) Problem definition (b) Model 1: Structural

response analysis (c) Model 2: SDOF Analysis

Figure 6.7 Equipment models for developing single degree of freedom models [6]

6.3.11 Procedure for accounting for deck accelerations in the SDOF analysis

The deck is deflected and accelerated principally by the vertical component of the loading. The deck accelerations will cause inertia loads in the piping and piping supports, which must be taken into account. In addition, piping with multi-point attachments will be subjected to differential displacements at each of its supports. Deck accelerations may be determined using SDOF analysis using the following procedure:

1. Identify reference points where the deflection / acceleration need to be determined;

2. Apply a unit differential pressure to Model 1 in Figure 6.7. Note that Model 1 should include all parts of the structure where deflection reference points identified in 1 are needed;

3. Determine the deflection / acceleration at the reference points corresponding to a unit differential pressure;

4. Determine natural period of the deck for vertical movement;

5. Determine equivalent mass and stiffness of SDOF model (Model 2 in Figure 6.7);

6. The mass is considered the total mass of the deck and the equipment multiplied by the load mass factor KLM, which may be determined from tables in the Interim Guidance Notes [2];

7. The stiffness is then determined such that the mass determined in step (6) above will have the same deflection under an applied load equal to the unit load multiplied by the deck area;

8. The load is a force time history equal to the linearised deck pressure impulse multiplied by the deck area; and

9. Apply the load from step (6) to the SDOF system defined in step (5) to determine the displacements and accelerations at the reference support points.

6.3.12 Procedure for accounting for piping ductility in SDOF analysis

Plastic deformation of the piping or the piping supports will absorb the dynamic energy, and will lead to a significant reduction in the dynamic load factors (DLF and RDLF). Ductility can be accounted for by modifying the stiffness to be used in the SDOF model (Model 2 in Figure 6.7). The stiffness can be modified in one of the following ways:

Spring stiffness (k)

Equivalent mass (M)

F(t)


96

1. Use an equivalent stiffness of Rp /del; and

2. Use the three part curve shown in Figure 6.8.

Allowable deflection limit is often expressed in terms of a ductility ratio defined as del/dp. The Interim Guidance Notes [2] provide charts with dynamic load factors (DLFs) for various ductility ratios. Care must be taken to ensure that any weak members or connections will not fail before the deflection corresponding to the ductility ratio under consideration is reached.

6.3.13 Multiple Degree of Freedom (MDOF) Analysis

Where a structure cannot be idealised as SDOF, the only general method for determining its dynamic response to explosion loading is by Finite Element Analysis. In such cases the piping system together with the support structure should be modelled. Either decoupled or coupled analysis should be carried out. If the piping system is considered to have a significant effect on the topside structure then a coupled analysis should be carried out (e.g. this is usually done in the case of risers). In case the piping effect on the system can be simply represented by an additional mass then a decoupled analysis may be carried out. If the period of vibration of the piping system and the topside structure are of sufficient proximity such that resonance may occur, then a coupled analysis should be carried out.

Rp

d

R

Re

Res

ista

nce

Deflectiondel dm

Figure 6.8 Idealisation of non-linear load deflection characteristics

6.4 Acceptance criteria The main acceptance criteria that are usually considered for piping systems are:

• Strength limit • Strain limit • Deformation limit • Ductility limit

6.4.1 Strength Limit Failure is defined as occurring when the design load or load effects exceed the design strength. The criterion may be applied in the elastic as well as plastic regimes. Modified factors for loading and strength may be adopted to account for the fact that the blast is an extreme event.

6.4.2 Strain Limit Strain rupture limits under high strain rates are available for stainless steels. For other steel types approximate values have to be used. Table 3.6 of the Interim Guidance Notes provides strain limits for different classes of steel sections.

6.4.3 Deformation Limit This criterion reflects the fact that under large deflections the pipe may rupture from its support or the point of connection to other equipment.

6.4.4 Ductility limit A minimum ductility limit may be set to a ductile response under blast loading. Table 3.7 of the Interim Guidance Notes provides ductility ratios based on strain limits.

Figure 6.9 [63] shows typical escalation loads to equipment and to structure, where it can be seen that escalation to equipment occurs at much lower loads than escalation to structures.



Figure 6.9 Comparison of typical ‘escalation loads’ to equipment and structure [63]

1.00E-05

1.00E-04

1.00E-03

1.00E-02

0.01 0.1 1 10

Estimated pressure (bar)

Freq

uenc

y

Low er ALARP

AcceptanceEscalationto equipment

Escalationto area



7. FIRE LOADS

7.1 Introduction This Section provides a background discussion on fire loading and presents various methods for the determination of heat fluxes.

The Section starts with introducing a number of factors that will determine the fire loading. These include:

• Fuel type / system inventory (Section 7.1.1);

• Fire type (Section 7.1.2).

Section 7.1.3 provides examples of various fire types resulting in different total heat fluxes. The proportionality concept for determining heat fluxes is briefly discussed in Section 7.1.4.

Section 7.2 presents the fire characteristics that should be considered in the determination of fire loads. Section 7.3 discusses three methods for determining heat flux values (7.3.1 to 7.3.3); and Section 7.3.4 shows how the flame temperature corresponding to a particular heat flux may be obtained. Section 7.4 briefly discusses interaction of fires and explosions; and the combinations of analysis types for each hazard. 7.1.1 Fuel type Pressurised process plant may contain:

• Non-flammable gas (e.g. air); • Flammable gas (e.g. natural gas); • Pressure liquefied flammable gas (non-self

reactive e.g. propane, or self reactive e.g. butadiene);

• Pressurised flammable liquid (non-self reactive e.g. multi-component hydrocarbon mixtures, or self reactive e.g. propylene oxide); or

• Pressure liquefied toxic gas (e.g. chlorine and ammonia).

The inventory may be under high or low pressure. It should be recognised that process plants with a low pressure can become more pressurised in a fire due to the heating-up, boiling and thermal expansion of the contents, and / or if they contain:

• Low boiling point liquid (non-self reactive e.g. pentane, or self reactive e.g. acrolein);

• High boiling point self reactive liquid (e.g. organic peroxides);

• Cryogenic liquid (e.g. liquefied natural gas); or

• Solid self reactive (e.g. low melting point organic peroxides).

If non-self reactive liquid is heated in a vessel, its temperature will increase until it boils, bubbles form in the liquid and the liquid level rises / swells. If there is sufficient ullage space, the liquid level will not reach a vent fitted to the top of the vessel and only vapour will be vented from the vessel. However, if there is insufficient ullage space for the level swell, a two-phase mixture will be vented. Pressure relief systems are generally designed for vapour-only flow. Two-phase flow due to level swell is more likely during emergency depressurisation, which will cause flashing of the liquid in the vessel. The likelihood of liquid carryover is dependent on the system configuration, liquid level and liquid characteristics (e.g. foaminess).

If a self reactive liquid (i.e. a liquid capable of exothermic decomposition or polymerisation) is heated in a fire, a runaway exothermic reaction could be initiated before or on reaching the boiling temperature. In this case, there will be a very rapid increase in temperature since the reaction rate will double for every 10 to 15oC temperature rise. A very rapid increase in pressure may also occur if large amounts of gas / vapour are produced by the reaction.

7.1.2 Fire types The types of fire that may occur in onshore and offshore process or storage may vary significantly. The intensity of the same fire type may be different between an offshore and onshore facility. However, both could be subjected to fires fed by their own inventory, or from diverse nearby sources. Typically, a failure and leak is associated with pipework, flanges and small-bore fittings, such as instrument tubing. In the case of larger dimension pipework, full-bore failure is less foreseeable event and is more likely to be from a crack or partial failure of the gasket integrity in a flanged joint. A full bore failure might however occur from a failed bolted joint (due to fire or overload in explosion) or a pipe burst due to the combined effect of internal pressure and material weakening in the fire.



Ignited spills of flammable liquid will typically result in a pool fire. In the event of a major leak and where pipework, vessels or storage tanks containing such materials are within a bund, the fire may spread to the limit of the area and result in fire engulfment of the pipework / vessels / storage tanks in the bund. Plant and equipment outside the bund will be subject to radiation from the fire and possibly flame impingement. Where there is no bund or confinement, the pool fire will spread to its maximum extent and may impact on plant and equipment over a considerable distance. The topography, including any features provided to direct spillages away from the incident plant, e.g. sloped ground, will significantly influence the nature of the fire and its location. Since the pool is developed to its maximum size, there is very little depth associated with such an extensive pool and the fire rapidly burns out unless continuously fed with fuel. Hence, the duration of an unlimited pool fire is generally limited to the duration of the release from the inventory. For a bunded area, the duration can be significantly longer.

Jet fires burn as the pressurised inventory is released, and hence the duration is determined by the release rate and the mass of the inventory. Free (unobstructed) jet fires are highly directional so whether the target pipe is impinged by the flame is dependent on the geometry of the release. An unobstructed jet flame has an approximately conical shape with uplift due to flame buoyancy. More realistically, the fire will be obstructed by equipment, pipework and structures. In these circumstances, the flame will “splash” and distort from the cone shape. A degree of judgement is required in determining whether a given pipe or other plant item will be directly impinged by the flames.

For onshore plant, similar considerations may be made based on the extent of the process area in which the release is occurring. However, equipment density tends to be lower at onshore process plants, with an associated reduction in interference with the flame geometry.

The aim of these guidelines is to address the design aspects of protecting pipework against fires to mitigate the potentially catastrophic effect on surrounding plant and process equipment. It is a fundamental requirement to understand the time to failure in order to be able to design the surrounding facilities and

structures and to allow for the safe escape and evacuation of personnel. Careful thought also needs to be given to the design of passive and active fire protection systems, which can be critical in obtaining the extra time necessary to delay and ideally prevent the potentially catastrophic failure of the plant.

7.1.3 Heat loads Different fire types will result in different total heat fluxes, comprising radiative and convective components. For example:

• High pressure releases of fuel with a significant gas content will tend to produce a high convective heat flux;

• Pool fires and jet fires of liquid fuels tend to have a low convective heat flux;

• Higher hydrocarbons tend to produce more radiative flames;

• Ventilation-controlled fires in enclosed spaces can result in higher flame temperatures and higher heat fluxes due to restricted heat losses. Also, the heat is re-radiated from the surrounding walls into the fire, which supports the combustion process, resulting in very high temperatures;

• Ventilation-controlled fires in enclosed spaces can also result in lower heat fluxes if the air supply rate is too low; and

• Very large fires can produce higher flame temperatures and hence higher heat fluxes.

An ignited release of flammable fluids can result in a wide spectrum of fires with differing heat fluxes depending on the various factors outlined above.

In order to predict the heat up of pipework and pipework supports subjected to fire, it is necessary to understand and quantify the thermal load imposed by fires. This thermal load is a combination of convection from the hot combustion products passing over the object surface and radiation emitted by the flame to the object surface. In reality this is a complex process and following issues are relevant:

• The total heat loads will vary depending on the fuel type, the size and shape of the object and the location of the object within the fire;



• The relative proportion of radiative and convective load from a flame will vary depending on the fuel type and location of the object within the flame;

• The heat loads will vary over the surface of the objects; and

• The heat absorbed by the object will vary with time.

For each of the likely fire hazards identified, or for the worst foreseeable case, the fire load on the pipework of interest should be quantified and the appropriate heat load passed to the relevant disciplines involved.

7.1.4 Proportionality concept The concept of proportionality discussed in Section 3.2 applies to both fire and explosion loading. Proportionality must be considered in at least two aspects of the safety assessment of fire loading:

• The robustness of the risk assessment used; and

• The depth of the ALARP demonstration.

Figure 7.1 shows the fire hazard assessment framework, based on a flowchart produced by Cook and Phelp [64]. The determination of fire loading comes after the identification of potentially flammable releases and their frequency, and it is a prerequisite to determining the response of equipment, pipework and

structures to an ensuing fire. The boxes with dark shading are those tasks where the piping is expected to play an important role in the hazard management process.

7.2 Fire characteristics to be considered

The fire characteristics that should be considered in the determination of fire loads include:

• Whether the fire is a pool fire or a jet fire, confined or unconfined;

• Whether the fire is fuel- or ventilation-controlled;

• Composition of fire fuel and whether the fuel is one- or two-phase;

• Gas to oil ratio (GOR) in the burning fluid; • Whether the flame is obstructed or

unobstructed. Most of fires in offshore modules and land-based plants / process buildings will be obstructed by equipment, pipework and structures;

• For an offshore module, the size of the release and resulting fire with respect to the size of the module;

• Temporal and spatial variation of heat flux within a flame. As shown in [3] and [2], where a flame impinges on or engulfs an object, such as a pipe, the heat flux tails-off rapidly in the direction away from the flame footprint on the pipe [2, 3];

Figure 7.1 Fire Hazard Management Diagram [64]



• A small fire may impinge locally onto a target (such as pipe, item of equipment or structure), but will have little effect on surrounding pipework / equipment / structures. A medium size fire in the open may impinge on the target whilst the tail ends of the flame will impose a reduced heat flux on equipment / pipework and structures in close proximity to the target;

• A medium size fire confined in a compartment or offshore module is likely to impinge on a target and the combustion products will also fill the module. This will result in concentrated (sometimes also called “point”) and background loads of different respective magnitudes. The concentrated load will be imposed locally on part of a pipe, whilst the background load may affect all hydrocarbon carrying equipment within the module;

• In scenarios of ventilation-controlled confined fires within compartments or modules where (a) there is enough oxygen supply to the combustion process, and (b) where the fire is large in comparison with the size of the module, the heat flux may reflect from the module walls and enhance the combustion process. Such conditions result in very high heat fluxes;

• A medium and large fuel-controlled unconfined fire (in the open) in a land-based process plant or on an offshore platform will impinge or engulf targets in the path of its flame, resulting in a high point heat load. It will exhibit a lower heat load onto targets in the vicinity of the impinged / engulfed target; and

• The heat flux outside the flame rapidly reduces with the distance away from the flame.

7.3 Determination of heat fluxes

The heat flux imposed on a piping system may be determined using the following methods:

• Heat flux values tabulated in various standards and guidance, depending on fire type;

• Empirical and phenomenological models; and

• Field models (Computational Fluid Dynamics (CFD) models).

The tabulated heat fluxes have been obtained from measurements.

CFD models are generally applicable. However, due to the uncertainties associated with the prediction of the combustion process, the models should be verified for the problem in hand.

7.3.1 Tabulated heat flux values Several documents provide values for total incident heat flux for various hydrocarbon fire types in the offshore industry: The Interim Guidance Notes [2], Fire and Blast Engineering Project Phase II [3], The NORSOK Standard for Technical Safety [65], Statoil / Norsk Hydro / Scandpower guidelines [9] and the Institute of Petroleum guidance [10]. As more tests are carried out these values tend to change to reflect the latest data.

Data for heat fluxes from LNG, LPG, butane, gasoline and kerosene fires typical for land-based petrochemical plant can be obtained from the SFPE Handbook [66].

Table 3.10 summarises the heat fluxes from various fire types, as reflected in the latest studies. The tabulated values are for impinging or engulfing fires. Reference [3] shows that the areas immediately outside the footprint of the flame on the target pipe also receives heat directly from the flame, however, the heat load is lower than within the footprint. This indicates the spatial variation of heat load for flame obstructed by pipework, equipment and structures.

The definitions of the terms that are used in the Table for the purpose of the fire loading are as follows:

Maximum average load

A heat flux from a ventilation-controlled confined jet fire within compartment or module where (a) there is air supply to the combustion process, and (b) where the fire is large in comparison with the size of the module, so that the flame may fill it or the heat flux may reflect from the module walls and enhance the combustion process. Such conditions result in very high heat fluxes.

Maximum point load

A heat flux within the area of a flame footprint impinging on or



engulfing a target.

Background load

A heat flux within a module caused by hot products of the combustion process outside the flame.

The shape of a jet flame is normally represented as a cone or cylinder. The flame from a pool fire is usually considered as cylindrical. The flame footprint on the target is approximately equal to the flame diameter at the distance of the target.

The above terms describe the extent of fire loading, which result from the consideration of the size of the fire with respect to the size of the target pipe. As for example:

A large hydrocarbon pool or jet fire may engulf the whole piping system, where the latter is small relative to the engulfing flame. In this case the fire load will be the maximum average or point load evenly distributed over the whole surface area of the piping system.

Another example – a large pipe is impinged by a relatively small fuel-controlled hydrocarbon jet flame. In this case the fire load will be the maximum point load of 250kW/m2 evenly distributed around the circumference of the pipe and along its length, where the length will be equal to the diameter of the flame footprint on the pipe.

Yet another example – a piping system in an enclosed module attacked by a large ventilation-controlled hydrocarbon jet flame. In this case the whole piping system will be engulfed by the jet flame with the maximum average load of 400kW/m2.

The relationship between the size of pool of LNG, LPG, butane, gasoline and kerosene, and their respective fire loads is addressed in the SFPE Handbook [66].

The Stefan-Boltzman relationship may be used to obtain the flame temperature from the heat flux and vice versa.

The following observation may be drawn on the recent findings on heat flux determination [8]:

• On the whole, the information given in the Interim Guidance Notes [2] remains valid for all the fire scenarios which concern jet fires and pool fires in the open, although it is worth noting that new information is now available in the case of two-phase jet fires. The information given for jet fires and pool fires in a module is somewhat outdated in light of recent advances in experimental and theoretical work.

• Unconfined Two-Phase Jet Fires: The incident total heat fluxes (radiative and convective) measured on the pipe target were significantly higher for the mixed fuel tests than for the crude oil only tests, by a factor two in many cases. Typical values were in the range 50kW/m² to 400kW/m².

7.3.2 Empirical and phenomenological models

These have been discussed in the Phase 1 [1] Fire Loading reports and the Interim Guidance Notes [2]. In these studies it was generally accepted that the semi-empirical models provide the most accurate and reliable prediction of the physical hazards associated with fires, providing that they are being applied within the validated limits of the model.

7.3.3 CFD models CFD models are the only models that can represent the fluid dynamic processes within the flame and enclosure. They are developing rapidly and are expected to improve in accuracy over the coming years [8].



Table 7.1 Heat flux values for various fire types (kW/m2)

Fire Loading (kW/m2)

Type of Fire

Maximum Average

Load

Maximum Point Load

Background Load

Pool fire crude small fire open or enclosed area fuel or ventilation controlled

n/a

150

n/a

Pool fire crude medium or large open area fuel controlled

n/a

150

n/a

Pool fire [9] crude medium or large enclosed area fuel controlled

n/a

1501

1002

Pool fire LNG open area fuel controlled

n/a

40 – 180

(depending on pool size [66])

n/a

Pool fire LPG open area fuel controlled

n/a

40 – 120


n/a

Pool fire Butane, gasoline, kerosene open area fuel controlled

n/a

130 – 20


n/a

Jet fire crude medium or large enclosed area ventilation controlled

400

n/a

n/a

Jet fire crude small open or enclosed area ventilation controlled

n/a

250

n/a

Jet fire [9] crude leak rate>2kg/s enclosed area fuel Controlled

n/a 3501 1002

Jet fire [9] crude leak rate>0.1kg/s enclosed area fuel Controlled

n/a 2501 0

Blowout jet fire crude or gas (always considered as large) open area fuel controlled

n/a

300

n/a

Blowout jet fire crude or gas (always considered as large) enclosed area fuel controlled

400

n/a

n/a

Notes: 1. Heat flux for the determination of material properties, 2.Heat flux for the determination of thermal expansion and boiling of process segment contents.



7.3.4 Derivation of temperature from heat fluxes

The fire loading acting on pipework, equipment or structure is expressed in terms of heat flux or a flame temperature and heat transfer coefficient. The flame temperature equivalent to a heat flux can be obtained from the Stefan-Boltzman relationship

4fTq σ= where

q is the heat flux (W/m2)

σ is the Stephan - Boltzman constant:

= 5.6697 × 10-8 W/m2K4

Tf is the flame temperature

As can be seen from Table 7.1, one of the main factors affecting the thermal loading of piping systems is its location relative to the fire (impinged / engulfed or not impinged / engulfed). For impinging and engulfing flames the heat transfer mechanisms are by thermal radiation and convection and both of these mechanisms must be included in the prediction of the heat received by the target.

Objects outside the flame receive the heat by thermal radiation. It should be recognised that in addition to receiving radiation from the hot flame, a non-impinged/non-engulfed surface may also receive re-radiation from other surfaces. In certain instances, e.g. compartment fires as mentioned in Section 7.1.2, the re-radiation component may considerably increase the total received flux.

The values of heat transfer coefficients are addressed in Section 8.

7.3.5 Probabilistic fire simulation As it may been seen from previous Sections, the methods currently used for the prediction of fire loads are deterministic. However, fire scenarios and characteristics are of such a nature that can be expressed in a probabilistic manner. A planned joint industry project will make an attempt to develop procedures for probabilistic fire simulation [66].

7.4 Interaction of fires and explosions

Figure 7.2 presents a typical chain of explosion and fire events after an accidental release of flammable fluid. It can be seen that a release of hazardous gas or liquid can lead either to 1. no ignition, 2. immediate ignition and subsequent fire or 3. formation of a combustible cloud, which then may lead to a delayed ignition and gas explosion. As the initial explosion and fire events develop they may escalate to new explosions and fires.

Considering the risk, fires have a higher occurrence frequency than explosions, but explosions have more severe consequences. However, it is also possible that severe consequences develop from severe fires where the latter occur at a much lower frequency. Indeed, fatalities often arise due to smoke inhalation from fire and not from overpressure loading due to explosions.

Clearly the process of escalation has to stop as early as possible in the chain, which has to be confirmed by very low outcome frequency of risk at that point.

7.4.1 Combination of methods used for determining the explosion and fire response

When considering the combined effect of fires and explosions, any combinations of analysis types is permissible depending on the level of risk for fires and explosions. It is possible to combine:

• A simplified treatment of explosions with a simplified treatment of fires;

• A simplified treatment of fires with an advanced treatment of explosions; or

• A simplified treatment of explosions with an advanced treatment of fires.

• An advanced treatment of fires with an advanced treatment of explosions.

However using an advanced method for fire and a simplified method for explosions (or vice versa) should be compatible with the principle of proportionality of risk discussed in detail in Sections 3.1 and 3.2.



Escalation to fire

Escalation to explosion

Escalation to fire


Escalation to fire


Escalation to fire


Escalation to fire


Jet fire

Explosion

Jet fire

Jet fire and pool fire

Pool fire

Immediate ignition

Delayed ignition

No ignition

No ignition

Ignition

Formation of flammable cloud

Flammable gas

Flammable liquid

Release of flammable fluid

Figure 7.2 Typical chain of potential events following the release of flammable fluid



8. DESIGN OF PIPING AGAINST FIRES

8.1 Introduction Process equipment and pipework have a much broader variety of response to fires than structures. The response ranges from the simple sagging of a dry pipe to the potential catastrophic explosion of a pressure vessel, BLEVE, or a rupture of hydrocarbon-transporting pipe.

This Section considers the techniques that may be used to determine the effects of fire loads on piping systems. The pre-requisite of design and protection is the knowledge of the processes that one designs for and protects against. Section 8.2 describes the behaviour and failure modes of piping systems and their components affected by fire. Section 8.3 discusses the design of pressurised piping systems. Section 8.4 presents a discussion on the design of water deluge and other fire safety systems. Section 8.5 provides a brief discussion on the response of piping supports subjected to fire. Section 8.6 discusses the response of flanges, bolts and welds to fire. Section 8.7 addresses passive and active fire protection systems. Finally, Section 8.8 discusses how fire protection systems may be optimised during the design process.

The design method is illustrated by an Example in Appendix 2 and this will be referred to as Example 2.

8.2 Effects of fire on pipework The response of pressurised piping systems to fire loads is very variable. The main considerations are:

• Fire scenarios that the piping systems and connected vessels are exposed to;

• The size of the pipe (outer diameter); • Passive fire protection (if provided); • Thermal insulation; • Material of construction and thickness; and • Pipe contents.

The heat is transferred from the flame onto the pipe outer surface. It is then conducted through any protection coating and insulation material, through the pipe wall and into the fluid inside

the pipe. This results in the temperature rise of the fluid, its boiling and evaporation.

In general, critical piping systems should have passive fire protection (PFP) coatings. However, as recommended by the NORSOK Standard [45], and the guidance by the Institute of Petroleum [10] and Statoil / Norsk Hydro / Scandpower [9], the application of PFP should be seen in the wider context of a safety plan for protection of piping against fires.

If a process line is partially or completely insulated for process reasons (e.g. heat conservation) it may perform well under fire loads, but some lagging materials and the sheeting that hold them in place are unlikely to be effective in a fire. One should seek a report from a formal fire test before credit is taken for the fire resistance of thermal insulation.

The main material types are carbon steel, lined carbon steel, stainless steel, GRP and Kunifer. These materials have different thermal and mechanical properties at elevated temperatures, and will behave differently under fire loading conditions. The required material properties are linked to the function of the pipe itself and so evaluation should be carried out on a system-by-system basis.

The thermodynamic behaviour of the pipe contents needs to be taken into consideration. Flowing liquid inside the pipe will be able to remove local heating when the flow velocity is relatively high. One of the first responses to an accidental fire is, however, plant emergency shutdown and a pipe with a non-flowing fluid should therefore be considered as the realistic case.

8.2.1 Heat flux received by the target pipe

Fire characteristics are addressed in Sections 7.2 and 7.3. This Section addresses the heat received by the pipe.

Heat transferred from flame onto the pipe surface or onto the coated surface For an impinging flame, the heat transfer onto a pipe is by radiation and convection. The flame heat release rate and the temperature distribution



within the flame are influenced by the fuel, supply of oxygen and soot content. The radiation component is higher when soot is in high temperature areas. Soot outside high temperature areas produces dense smoke and often also steam, which reduce the radiation component. The convection component is higher for high velocity flame than for low velocity flame. For jet flames, the flame velocity depends on the size of the leak and pressure within the leak source. The velocity of a pool fire flame is approximately constant with respect to the height from the pool surface. The heat transferred to the pipe reduces with the rising temperature of the pipe surface.

Heat transferred through pipe fire protection coating or thermal insulation The heat transfer mechanism through the protective coating or heat conservation insulation is mainly conduction. The thermal and heat transfer properties vary with temperature for both cases.

8.2.2 Effects on the pipe Figure 8.1 illustrates a cross-section of a pipe, which carries a 3-phase fluid, say a mixture of hydrocarbon liquid, gas and water. The outer surface of the pipe or the pipe fire protection coating, thermal insulation, or protective shield receive heat from the fire. The following time-dependent processes take place [9, 10, 67]:

• Heat conduction through the pipe wall with resulting temperature Ts (outer surface temperature), T1, T2, T3, etc. and Tsi (inner surface temperature);

• Heat transfer from the inner pipe surface to the pipe contents (qil to the liquid and qis to gas);

• Thermodynamics of the pipe contents (the temperature and mass of liquid Tl and ρl, the temperature and mass of gas Tg and ρg, evaporation of liquid mlg and condensation of gas mgl);

• Variation of pressure in the pipe due to depressurisation, counter-acted by the increase of the pressure due to evaporation, boiling and expansion of vessel contents.

• Strains and stress in the pipe wall; • Thermodynamics in the depressurisation

pipework, which may experience low

temperature due to a sudden change of pressure;

Tm

To

Tg

Tl

qo

qig

qil

lgm•

tm•

gρ

ρl

Ts

Tsi

T1,T2,T3

glm•

Figure 8.1 An illustration of processes in a

pipe carrying a 3-phase fluid

• Low temperature leading to possible material embrittlement of the depressurisation valve and pipework resulting from the rapidly expanding stream of gas through the blowdown valve; and

• Variation of thermal and mechanical properties of the construction materials with temperature.

The above processes are closely coupled together and influence each other.

Heat conduction through the pipe wall Heat is then conducted through the pipe wall to the pipe contents. Both specific heat and thermal conductivity of the pipe material are temperature dependent and they are different for various construction materials [8]. Temperature gradients exist in the radial, axial and circumferential directions. As the heat transfer values are higher for liquid contents, the temperature of the inner pipe surface will be lower in liquid space than in gas space.

Figure 8.2 illustrates the temperature gradients in the radial and circumferential directions at various time instances after the start of an impinging fire. Note the reducing extent of the bottom colder surface of the pipe due to the evaporation of the liquid.



Figure 8.2 Illustration of the temperature distributions in the radial and circumferential

directions of a pipe cross section

Heat transfer from the inner surface of the pipe wall to the pipe contents The mechanisms of heat transfer from the inner surface of the pipe wall to the contents are by radiation and convection. The convection component is heavily dependent on the temperature of the pipe inner surface. The convection process rises moderately with temperature until a nucleate boiling state is reached at the pipe inner surface, when the heat transfer is very high. However, the heat transfer reduces when the nucleate boiling changes to film boiling.

In the nucleate boiling regime, bubbles are formed at a number of areas on the (heated) inner pipe wall surface. As soon as the bubbles are detached from the surface, they are dissipated in the liquid. As the temperature of the inner wall surface rises further, the number of nucleation areas increases and the bubble generation rate is so high that continuous columns of vapour appear. As a result, very high heat fluxes are transferred into the liquid in this region. The heat flux increases rapidly with increasing temperature difference between the wall surface and the liquid until a peak heat flux transfer is reached [68].

After the peak heat flux transfer is reached, any further increase in the temperature difference causes a reduction in the transferred heat flux. The reason for this phenomenon is the blanketing of the wall surface with vapour film that restricts liquid flow to the surface and has a low thermal conductivity. To begin with, this

film boiling region is unstable, collapsing and reforming under the influence of convective currents and the surface tension. The transferred heat flux decreases as the surface temperature increases, because the average wetted area of the heated wall surface decreases. Then the transferred heat flux drops to a minimum, because a continuous vapour film covers the heated surface. Thereafter the transferred heat flux begins to increase as the temperature difference increases, because the temperature of the heated wall surface is sufficiently high for thermal radiation effects to augment heat transfer through the vapour film.

Thermodynamics of the pipe contents The thermodynamic behaviour of pipe contents includes evaporation, boiling, vapour expansion and pressure, as the liquid and vapours / gas are progressively heated by the pipe shell. Heat and mass transfer takes place between the fluid phases (which are typically multi-component hydrocarbon liquid, multi-component hydrocarbon gas, water and steam) through condensation and vaporisation.

When there is a flow in the pipe, the fluids inside the pipe carries heat away from the fire-affected area. For practical purposes, however, this may be ignored as one of the first plant responses to an accidental fire is shutdown, which stops the fluid flow within a minute.

Emergency depressurisation Emergency depressurisation of pipework may be carried out through the activation of a blowdown



system, process safety valves (PSV) or both. Normally, blowdown systems are used for a rapid depressurisation in emergency, whilst PSVs are used for pressure relief to control process conditions.

During depressurisation, pressure in the pipe reduces, depending on the thermodynamic conditions of the pipe, the pipe contents and the orifice size. This is counter-acted by the pressure rise due to the heating-up of the pipe contents, their expansion, evaporation, boiling and the expansion of vapours or gas in the pipe.

As a result of the activation of depressurisation (blowdown), the rapidly expanding stream of gas through the blowdown valve causes a temperature drop in the valve and blowdown pipe, which may reach sub-zero levels.

Stress and strain in the pipe shell Three dimensional stress / strain state occurs in the pipe shell due to:

• Applied internal pressure in the pipe acting in all directions, including the longitudinal direction due the closed pipework system (end-cap action).

• Temperature gradients in the radial direction within the pipe wall (especially where the liquid inside the pipe is in contact with the pipe wall) and in the circumferential direction between the liquid and vapour space.

• Longitudinal temperature gradients between the pipe part exposed and unexposed to fire, and due to pipe supports that limit the thermal expansion of the pipe.

• External loads through pipe supports caused by deformation of the fire-affected structure the pipe is connected to.

• Stress concentrations due to bends, changes of pipe thickness, attachments, etc. Stress concentrations caused by changes of stiffness may vary with time depending on the temperature rise of the object causing the stress concentration, as it may lose stiffness at high temperature.

8.2.3 Failure mechanisms of pipework Failure mechanisms that cause the loss of pipework functionality, i.e. loss of containment, are due the effects of excessive stress and differential thermal expansion.

Pipes without PFP

Based on excessive applied stress compared with the material yield stress or the Ultimate Tensile Strength (UTS) at high temperature of the pipe, the pipe may fracture. The fracture may have a high stress concentration at its tip, which may cause unzipping of a part of the pipe. However, it should be noted that at elevated temperatures, the material ductility is enhanced so that the failure mode of the pipe is likely to be ductile rupture.

Based on observations of process plant accidents and pipe tests, pipes seldom explode. Failure is usually found to occur in connections.

Unprotected pipe affected by fire will rapidly heat-up and expand, where the expansion will be resisted by pipe supports and equipment that the pipe is attached to. This may result in excessive permanent deformation and rupture of the pipe.

The rapidly expanding stream of gas through blowdown valve causes a temperature drop in the valve and blowdown pipe. This may cause embrittlement of the valve and pipe material and their fracture.

Pipes with PFP The pipe being designed may have fire protection coating whilst a branch pipe attached to it may be unprotected. The unprotected pipe will rapidly heat-up from the fire and conduct the heat into the protected pipe. A high temperature hot spot may thus develop, which may lead to the applied stress to locally exceed the yield stress or UTS leading to fracture. This is illustrated in Figure 8.3.



Figure 8.3 Typical temperature distribution in the intersection of a large fire-protected and a small unprotected pipe

The temperature drop in the blowdown piping described for the pipes with no PFP also occurs in protected pipes.

Pipe supports Pipe supports are normally connected to primary or secondary structural steel or pipe racks. Excessive deformation during fire of the primary and secondary steel, and pipe racks would lead to excessive movements of pipe supports, where the latter may cause pipe rupture or loss of integrity of flanges.

Excessive thermal expansion of unprotected pipe may lead to rupture of its supports.

Flanges Ordinary API flanges Flanges affected by fire develop differential thermal expansion, which alone or together with a low material strength capacity at high temperature may lead to loss of tightness.

Flange seals may be damaged due to high temperature leading to loss of tightness.

Based on [43 & 67], analyses of unprotected flanges show that the flange body heats up faster than bolts where the latter are shielded by the former. The nuts heat up faster than bolts because they are not shielded from the fire. The thermal expansion of the flange body together with the pre-tensioning of bolts and the loss of load carrying capacity of the nut threads at high temperature cause the over-stressing of the nut

threads and nuts shear off bolts within 5 minutes after the start of the fire. This was confirmed by measurements [42 & 43].

Reference [1] describes experiments to determine the thermal response of flange connections, the time to loss of tightness and failure modes during jet-fire attack. The tests established that the tightness of a flange connection may be lost and new leaks formed between 1 to 8 minutes after the start of the fire. An asymmetric temperature distribution develops in the flange connection even in an engulfing jet flame, the downstream side of the flange being hotter. The loss of tightness was attributed, in all cases, to the same cause viz: the decrease of the contact pressure because the temperature induced expansion of flange bolts was higher than that of the flanges. Moreover, because of the thermal gradient in the flange connections, the bolts elongate differently and the leaks occur in the areas with higher temperature. It was concluded from this work that:

• Standard tests according to API and British Standards provide no real information on the loss of tightness in real fire scenarios of jet-fire impingement;

• In the tests, the elongation of the bolts remained in the elastic range;

• The sealings showed little or no damage and, after cooling down at the end of the tests, some test samples even re-gained their tightness; and

• In a real fire case, the loss of tightness would lead to damage of the sealings as the leaks would ignite.

As part of the Commission of the European Community (CEC) funded joint project on hazard consequences of Jet fire Interactions with Vessels containing pressurised liquids [8], the Battelle Institute performed jet-flame impingement trials on unprotected and protected flange connections and found that:

• Typical LPG flange connections, and some new ones tested, do not resist jet fire attack;

• The time to loss of tightness depends on the intensity and position of the jet fire and can be as short as one minute;

• Standard API 92 and BSI 87 tests provide no real information about loss of tightness in a realistic jet-fire scenario; and



• New protective measures are required for jet fires.

Hubb connectors and clamp flanges Compared with API flanges, hub connectors and clamp flanges lose tightness more rapidly as for an unprotected flange the bolts are directly exposed to fire and lose their pre-tension and strength [43].

Compact flanges Experiments show that unprotected compact flanges respond to fire in a manner similar to API flanges. This with the exception that the loss of tightness of an API flange is a gradual process whilst the leak from the compact flange affected by fire comes suddenly.

Valves The failure mechanism of valves exposed to fire is normally differential thermal expansion leading to loss of tightness.

Based on [8], standard fire tests of valves may not represent the realistic fire that the valve may be exposed to.

Valve actuators Unprotected valve actuators may lose their functionality within few minutes after the start of the fire [43]. PFP protection to actuators needs to keep temperature rise down to that which can be tolerated by the internal cabling and instrumentation.

Valve – actuator assemblies A valve-actuator assembly where the valve is protected and the actuator is not protected normally leads to the situation where heat is conducted through the actuator to the valve. Within 15 minutes after the start of the fire this leads to such temperature differentials that will cause the valve leaking [43].

Valve seals Sealing devices in valves may break down at elevated temperatures.

Bolts Bolted connections will need to be examined as their performance in a fire could possible be short-lived. A bolt pre-tensioned at ambient temperatures will lose the pre-tensioning when heated due to thermal expansion and the change of modulus of elasticity at elevated temperature.

Safety valves The thermal expansion of a helical spring above its normal operating temperature will tend to reduce its tension and hence lower the set pressure, so the effect tends to be in a safe direction. On the other hand, unequal thermal expansion can distort the valve spindle leading to jamming of the valve and a possible reduction in the discharge capacity [8].

Bursting discs Similarly to safety valves, the effect of high temperature on a bursting disc is to lower the bursting pressure or destroy it altogether, and hence the tendency is in a safe direction [8].

8.3 Design of piping systems 8.3.1 Design process outline Figure 8.4 gives an outline of the design process for pressure systems. It provides a methodology for optimising the design for loss prevention until the risks are reduced to meet requirements. When working through the design process, it is important to take into account the total facility or plant design, and not just the pressure system in isolation.



Figure 8.4 Outline of design process for pressure systems against fire



The choice of alternative designs may be limited by process requirements. Consequently, it is important to establish at an early stage (e.g. during the conceptual design phase) what items are essentially fixed and therefore the options that are available for the remaining equipment. The operating data should be recorded in the process flow diagrams (PFD), equipment data sheets and the piping and instrumentation diagrams (P&IDs). A hazard register should also be compiled.

The elements of Figure 6.4 are described in the Sections below.

8.3.2 Fire protection philosophy General The fire protection philosophy should be established at the onset of the project based on operator’s risk criteria. As a minimum, this should include the following information:

• References to regulations, standards and guidance used for the project;

• Installation risk criteria; • Fire and explosion strategy (FES); • Overall strategy for prevention of

escalation; and • Overall, objectives of passive and active

fire protection.

The experience shows that the fire protection philosophy should be a central document for the project including both qualitative information and quantitative data, as sub-division of the installation into fire / explosion areas, requirements for endurance of safety barriers (in terms of time), fire scenarios and loads, etc. This information should be included as soon as it is generated in the steps described below. Safety barriers [37 & 69] are also called essential safety systems [26].

Offshore installations The overall strategy for prevention of escalation is typically that the facilities that are to fulfil main safety functions, such as escape, evacuation and rescue facilities, must be designed such that they survive and remain functional in the events of the frequency of 10-4 (one event in 10000 years). For environmental reasons, survival of the facility itself may also be a requirement [30].

This in turn dictates the design frequency and Dimensioning Accidental Loads (DAL) for which escalation to the escape, evacuation and rescue facilities must be prevented. This in turn dictates the fire and explosion loads that the explosion rated firewalls on the boundary of a fire / explosion area must successfully resist.

Simplistically, an isolateable process segment should ideally be entirely contained within one fire / explosion area. In order to prevent all escalation events, all pipework and their components should be fire-protected, which would not be economically viable. A limited controlled escalation may therefore be allowed by stating that, as for example, all pipework and equipment with flammable inventories within a fire / explosion area and process segment greater than a specified mass of flammable material requires fire protection. The mass to be specified could arise from the application of the general operator’s fire/explosion protection strategy.

These relatively simple rules can enable a controlled escalation to neighbouring equipment within a fire / explosion area. Escalation to neighbouring fire / explosion area will be prevented and escalation to the escape, evacuation and rescue facilities will also be prevented.

The fire protection philosophy should also specify the minimum required fire endurance for the escape, evacuation, rescue and Temporary Refuge facilities.

Onshore installations The consideration that the systems that carry out main safety functions, such as escape, evacuation and rescue facilities, must be designed such that they survive and remain functional in the events of the frequency of 10-4 (one event in 10000 years) can also be applied for onshore plants.

Several process segments normally form a plant unit and several units may form a plot of a refinery or chemical plant. Due to chemical processing reasons, equipment that belong to one process segment may be geographically located within or in proximity of another segment within the same plant unit, which may make it difficult to prevent escalation between process segments.



Also plant units may be in proximity of each other, however the risk of escalation from one unit to another should be minimised.

The distance between plant plots should be such that the risk of escalation from one plot to another is prevented. Typically, plant plots that process flammable substances should be in such a distance from tanks farms that an event originating in the process plant would not escalate to the tank farm and vice versa.

Operators’ internal standards may provide guidance on separation between plots and units, which should be reflected in the fire protection philosophy.

New installations For new projects, there should be a fire protection philosophy established at the project onset, or obtained from a previous phase of the project and developed further.

Modifications of existing installations The fire safety philosophy should also be prepared or obtained for modifications of existing installations and confirmed with the actual situation on the platform or on site. The confirmation and contact with the operator would alleviate potential problems at the later stage, where the philosophy and resulting design may not fit to the equipment actually installed.

Some operators may wish to use an approach based on the latest guidance for modifications of old plant that was originally designed based on out-of-date guidance. This difference may have impact on design solutions used for the modifications. It is advisable to bring the differences to the attention of the operator at an early stage in order to avoid mis-match in the plant, unnecessary cost, delays and embarrassment.

It should be noted that some very old plants may have immaculate track record with regard to safety although they were built to old standards. This also needs to be taken into consideration in the design of modifications. The key to a successful design in this case is to obtain adequate identification and understanding of the hazards that the old plant may have, how the old plant will operate and behave when the modifications are implemented, and how the

hazards associated with the old plant will interact with the modifications and vice versa.

Fire barriers One of the main objectives of a fire / explosion barrier is to minimise the risk of escalation between fire / explosion areas.

The main aim of this task is to divide the installation into fire areas and determine strategies for the prevention of escalation. Input to this activity is the information on general plant layout. All project disciplines should participate in the general layout development.

Boundaries of fire areas are normally the same as boundaries of explosion areas as explosion is often followed by fire. An explosion should not breach a firewall (on an offshore installation) and allow flame penetration into the neighbouring area.

In an onshore process plant this segregation is normally achieved by location of plant units and distances between them.

Hazardous area classification should also be taken into consideration in conjunction with this activity.

For modifications of an existing installation the designer should try to maintain the existing sub-division into fire / explosion areas. Experience shows that changing this may be a very costly exercise.

Determination of fire scenarios The fire scenario definition includes:

• Fire fuel; • Fire type;

pool fire vs. jet fire (gas or liquid droplets);

fire originating from process equipment vs. that from blowout;

fire on the platform / within a process plant unit or on the sea surface;

ventilation-controlled vs. fuel-controlled; • Fire loading (heat flux, and duration); and

• Presence of a deluge system.



The fire loading is preliminary at early design stage and subject to detailed information following the process design.

New installations The fire scenario determination for a new installation should start with the consideration of composition and flammability of the fire fuel. There are substances where there is adequate information available to do so. Examples of such fluids are offshore hydrocarbons, where heat fluxes are generally known, although information for heavier hydrocarbons may not be readily available.

The flammability and fire heat fluxes for fluids processed in petrochemical plants onshore may not all be publicly available and may have to be obtained from the plant operator.

Fire loads may be obtained in the form of heat fluxes or flame temperature using the methods described in Section 7. The Stefan-Boltzman relationship may be used to obtain the flame temperature from the heat flux and vice versa.

Modifications of existing installations In the offshore industry, hydrocarbon reservoir compositions change with time, and there are typically less hydrocarbons and more water in the well fluids in the later life of the reservoir. Such substances will be less flammable, and flame heat fluxes may be expected to be lower than for early life hydrocarbons. If this data is not available, fire loads for known hydrocarbon compositions should be used and it should be noted that the ignition probability will be lower for a high water content of the reservoir fluids.

It has to be assumed that the existing plant was originally designed for accidental fire and this fact should be included in the design, as potential accidents associated with modifications will have to interact with the existing plant and vice versa.

Influence of firewater Reference [3] shows that firewater deluge may reduce heat flux from both pool and jet fires if supplied at adequate rate, and over an adequate area.

Dedicated deluge to surfaces of process systems of which pipes form a part may provide adequate cooling in case of a pool fire if applied

before the surface affected by fire reaches 100oC. Above this temperature the deluge is unlikely to establish a continuous water film over the surface required for cooling.

A dedicated deluge will not provide surface cooling in case of a jet fire impinging on a pipe, as the jet momentum would push the water away from the pipe surface.

A firewater stream from a monitor would provide some cooling of pipe surface impinged by a jet flame, as the stream has both adequate momentum and water rate to penetrate the jet flame and establish a continuous water film on the pipe surface being cooled. This facility may be limited to smaller fires and smaller targets and one monitor may not be able to cover the whole surface.

8.3.3 Process design Equipment is selected as a part of the process design using process design guidance. This also includes designing against fires using the preliminary fire loading from the fire protection philosophy. The process design guidance and its background must be well understood and it must be confirmed that it covers the fires considered.

The plant being designed is sectionalised into isolateable process segments equipped with pressure relief valves (process safety valves (PSV)), emergency shutdown valves and, if appropriate, depressurisation (blowdown) valves.

The main function of PSVs is to relieve pressure in case of the plant experiencing upset conditions, although they also may play a role in emergency depressurisation.

The process segment is isolated by emergency shutdown valves (ESDVs), with closing time normally being within 60 seconds. ESDVs will normally shut the process segment off and isolate it from the connected process segments and / or equipment.

The aim of the depressurisation (blowdown) valves (BDVs) is to rapidly depressurise the contents of the process segment to vent or flare. This is carry out in some plants in accordance with a pre-set sequence (sequential depressurisation) to spread the flare / vent load.



The shutdown control of the plant is normally determined by means of Cause and Effect Charts, which relate the various pressure and temperature levels in the plant and the detection of accidental events to the level of shutdown. The plant shutdown levels typically comprise:

• Process shutdown; • Emergency shutdown; and • Abandonment shutdown.

8.3.4 Hazard evaluation and determination of criticality level

The level of design analysis is selected on the basis of the criticality level of the piping, which requires an evaluation of hazards associated with the piping operation. This is addressed in Section 3.3.

8.3.5 Detailed fire thermal loading The preliminary data in the fire protection philosophy are further developed to determine

fire loading. The methods for the determination of the detailed fire thermal loading are described in Section 7.

8.3.6 Fire response analysis of pipework

Figure 8.5 illustrates the process of assessing fire response of piping systems. The first step in the figure, which is to carry out an analysis for determining the pipe internal pressure is usually carried out using a multi-physics method, as described below.

This analysis is based on the IP [10] and Statoil, Norsk Hydro, Scandpower guidelines [9]. It involves the computer simulation of all the multi-physics processes described in Sections 8.2.1 and 8.2.2. This also satisfies the requirements of the NORSOK Standard for Technical Safety for rupture calculations [65]. The method is outlined in Figure 8.5, which provides further information on the calculation procedures for pipework in Figure 8.4.

Figure 8.5 Outline of the multi-physics design method for pipework protection against fire

Analysis of pipe internal pressure The following input data are required from the process and mechanical design:

• Fire types, size, duration, impinging / engulfing / non-impinging fire, fire loads and heat transfer properties;

• Process segment layout; • Location of isolating valves defining the

limits of the segment; • Volume of segment; • Pressure of segment;



• Fluid temperatures of segment; • Fluid composition of segment; • Wet and dry surface areas of segment; • Types and thermo-mechanical data of

construction materials of segment varying with temperature;

• Dimensions of components, including material thickness;

• ESD response time from the start of the fire;

• The size of blowdown orifice; • Backpressure in the depressuring system; • Depressuring pipe data: Diameter, wall

thickness and length; • PSV data: Orifice size, opening and closing

pressure and operation profile; • The capacity of the flare system; • Method for initiating depressurisation

(manual or automatic); • Time delay for initiation of

depressurisation; and • Acceptance criteria for failure. The analysis should start with no fire protection (or fire protection in accordance with operator’s policy) and minimum orifice size. This information can be changed and the analysis re-run should the acceptance criteria not be met, in which case the following additional data would be required: • Pipe cooling data: Applied deluge rate; • Data for fire protection coating: Material,

thickness, density, specific heat, thermal conductivity; and

• Data for thermal insulation (if appropriate for fire protection purposes): Material, thickness, density, specific heat, thermal conductivity, type of sheet metal cladding that holds the thermal insulation in place.

It should be confirmed by a full-scale test that credit can be taken for the thermal insulation. The thermal insulation must survive and remain functional throughout the fire to provide adequate protection.

The fire response analysis consists of the following two stages:

1. Calculation of the internal pressure in the pipe and pipe temperature, both varying

with time, are calculated in the first stage; and

2. The pipe stress response and whether the failure criterion is reached during the scenario.

Stage 1 – calculation of pipe internal pressure A time-dependent multi-physics analysis has to be carried out including all the processes described in Sections 8.2.1 and 8.2.2. For each time step in each process segment analysed, the following should be calculated:

• Heat received by the pipe; • Pressure in the system; • Temperature in all fluid phases; • Fluid composition in each phase; • Flow rate through the orifice; • Liquid levels; • Temperature of the construction material; • Temperature downstream of the orifice; and • Heat transfer at all interfaces.

In the calculation of the heat flux received by the pipe, it can be assumed that the flame and the object surface are diffuse grey bodies and that the ambient temperature of the surroundings is low compared to the flame temperature. Using these assumptions, in simplistic terms, the heat flux received by an object can be expressed as:

( ) ( )sfsffsconvradrec TThTTqqq −+−=+= 44εσε

where

εf, εs are the flame and surface emissivities;

σ is the Stephan - Boltzmann constant (5.6697 × 10-11kW/m2K4);

Tf,Ts are the flame and pipe surface temperatures (K);

h is the convective heat transfer coefficient (kW/m2K);

q rec is the heat flux received by the pipe (kW/m2K);

q rad is the radiative component of the heat flux received by the pipe (kW/m2); and

q conv is the convective component of the heat flux received by the pipe (kW/m2).



It is important to note that the amount of heat received by the pipe will reduce as the pipe heats up. Ts will increase whilst Tf will remain much the same.

The emissivities and convective heat transfer coefficient values can be obtained from Table 8.1 [10]. The higher the emissivity and convective heat transfer coefficient, the higher the heat input into the body.

Table 8.1 Typical parameters for hydrocarbon pool and jet fires

Fire type Description Open

pool fire

Large or confined pool fire

Open jet fire

Confined jet fire

Total incident flux (kW/m2)

50–150 100– 250 100–400 150–400

Radiative flux (kW/m2) 50–150 100–230 50–250 100–300

Convective flux (kW/m2) 0 0–20 50–150 50–100

Emissivity of flame 0.7–0.9 0.8–0.9 0.5–0.9 0.8–0.9

Temperature of flame (K)

1000–1400

1200–1450

1200–1500

1200–1600

Convective heat transfer coefficient (kW/m2K)

0 0–0.02 0.04–0.17 0.04–0.11

The emissivity of pipe surface depends mainly on the surface colour and its ability to reflect the heat. Light colour, polished or stainless steel surfaces have low emissivity, whilst the emissivity of grey surface is relatively high. However, a surface emissivity of 0.8 may be used for all practical surfaces and colours in petrochemical application, as the impinging flame will cause the paint to peel-off, the steel surface will oxidise and soot will darken the colour during the first minutes of the fire.

The starting pressure and temperature for the multi-physics analysis should be either the operating pressure and temperature, or the high-high pressure and corresponding temperature of the pressure system. The selected starting condition should be determined based on the operational / accident response philosophy of the plant.

As the heat is received by the pipe surface it is conducted and absorbed by the pipe material and transferred to the fluid inside the pipe. Full

thermodynamic behaviour of the fluid inside the pipe must be included. The fluid heats up, boils and evaporates and the vapours expand. The generation and thermal expansion of the vapours generate pressure that counter-acts the effects of depressurisation.

There may be situations whereby the fire only affects part of a process segment. In this case the whole affected shut-in fluid volume should be taken into consideration whilst the fire load may only be applied to a part of the segment surface.

Stage 2 – stress analysis The distribution of stresses in pipework is due to:

• Internal pressure; • Selfweight; • Weight of the fluid content of the pipe; • Weight imposed by flanges and valves; • Local stress raisers such as nozzles, elbows,

pipe branches, supports; • Thermal gradients through the pipe wall

thickness; • Thermal gradients in the pipe

circumferential direction caused by the interface between gas and liquid;

• Thermal gradients in the longitudinal direction if the pipe is only partially heated;

• Thermal gradients caused by different thickness of the material (different thermal mass) as thicker components heat up slower than thin ones;

• Thermal expansion; • Creep at stress levels close to material yield

stress and at high temperature; and • Boundary constraints, such as from the

supporting structure and vessels.

The production tolerances for pipe wall thickness allow for variation, which has to be taken into account. A widely used tolerance is +/-12.5%. The wall thickness should be taken into account when:

• Calculating the weight of the pipe, use the nominal thickness; and

• When determining the steel temperature of pipe, pressure profile and strength use the actual thickness, i.e. the nominal thickness



minus the production tolerances minus corrosion allowance.

The stress calculated in multi-physics analysis is normally limited to a 3D representation of stress in an infinite pipe. The account of stress raisers may be taken into consideration by the use of stress concentration factors. Stress concentration factors (SCF) express the stress rise due to local change of stiffness or load path. When a small pipe or nozzle is attached to a large pipe, the small pipe will heat up and lose its stiffness more rapidly than the large pipe and the effect of stress concentration from the attachment of the small pipe may be negligible.

In order to include most of the above bulleted effects of stress distributions the multi-physics analysis needs to be followed by a pipe stress analysis. The resulting internal pipe pressure and temperature calculated in the multi-physics approach are assessed and relevant load cases and material data for the relevant pipe temperature are transferred to the pipe stress analysis.

Level of Analysis There are four levels of analysis that may be applied in order of increasing complexity:

1. Section or Hp/A method. With this method it is assumed that the pressure inside the piping does not exceed the system design pressure at any time during the event. PFP is applied to the pipe to ensure that the temperature rise of the pipe does not lead to a reduction in the yield stress to below 0.67 of its ambient value. This is based on the fact that the operating safety factor at ambient temperatures in ASME B31.3 [11] is normally taken as 1.5, which is 1/0.67. For carbon steels the allowable temperature rise is usually taken to be 400°C. If the design pressure of the pipe is exceeded during the fire scenario, the allowable temperature rise for the specification of the PFP has to be reduced to ensure that the yield stress is not exceeded. Conversely, where pressures during the fire scenario remain substantially lower than the design pressure level, a higher allowable temperature rise may be used for the specification of the PFP;

2. Multi-physics method for straight pipe. In this method the temperature and pressure time curves are determined and used to

compute stress distributions and maximum stresses in the time domain. Failure is reached if the maximum applied stress exceeds the instantaneous (temperature dependent) yield stress during the fire scenario. This analysis utilises a geometrical model of an infinitely long pipe. For systems comprising pipes and vessels an analysis of each diameter/thickness combination is performed. The fluid layering in the pipe/vessel is accounted for in the stress analysis. Example 2 shows an application of this method;

3. Multi-physics method combined with elastic FE analysis. This method is similar to method 2 above except that snapshots of the temperature and pressure determined in the thermodynamics analysis are input into a finite element model which represents the detailed geometry of the pipe and includes the operating loads. Where a pipe element is used (e.g. Caesar II) stress concentration factors are incorporated as a database within the program and results are assessed as utilisation factors defined as the applied stress divided by the allowable stress. The allowable stresses being determined based on ASME B31.3 ‘code check approach’. An enhancement is to use shell type elements in the geometry model; and

4. Multi-physics method combined with nonlinear finite element analysis. In this method the results from the thermodynamic analysis (temperature and pressure time histories) are applied as input together with the operating loads to a finite element model of the piping system. The finite element analysis that is then carried out takes into account both geometric and material non linearities. Alternatively, the loading may be snapshots of maximum temperature and pressure as obtained from the thermodynamic analysis. Input would also need to include sets of stress strain curves at different temperatures. The failure criterion would need to be carefully selected. Safety factors can be implemented by stipulating conservative levels of allowable strain. However care is needed when utilising materials which have high yield to ultimate strain ratios at elevated temperatures.

An example of a level 3 analysis using a multi-physics approach (based on VessFire) / and how



it may be combined with a linear elastic approach (based on Caesar II) is provided in Example 2 in Appendix 2.

8.3.7 Acceptance criteria and evaluation of failure mode

General

• The method of calculating stress and determining acceptance criteria are vital to determine the time to failure and thus the remedial measures to be applied to prevent or delay failure. If the acceptance criteria have been met, the design will be adequate. Typically, these criteria are that:

• There has been no failure of the pressure system;

• The time to pipework failure exceeds the personnel evacuation time, which would include the time required to rescue injured personnel;

• The amount of hazardous material released at the loss of integrity is acceptable; and

• The consequences of failure are acceptable.

Piping systems The main acceptance criteria for piping systems maybe categorised under four broad categories, also used for the structural components:

• Strength limit; • Strain limit; • Deformation limit; and • Maintenance of structural and insulation

integrity.

Strength limit Where strength governs design, failure is defined as occurring when the design load or load effects, exceed the design strength (e.g. yield) in a manner that is similar to conventional design. The principal difference for fire resistant design is that modified factors on loading and/or strength may be adopted as it is an extreme event and the strength assessment must take account of the changes of mechanical properties with temperature.

Strain limit The following criteria should be considered when determining the strain limit to be used in design:

• Material used for the piping system; • Cross-sectional geometry and proportions;

and • The deformation capacity of any protection

material present.

Displacement limit The following criteria should be considered when determining the displacement limit to be used in design:

• Type of attachment (in terms of ductility) between piping and other equipment; and

• Type of support condition for piping system (whether it may be subjected to opposing displacements from support points).

Currently used criteria Stress Criteria The most widely used criterion is currently an assumption that failure occurs when the applied equivalent stress is equal to yield stress.

Conversely, the time to failure is time from the start of the fire to point when the applied equivalent stress becomes equal to yield stress.

It is the Von Mises equivalent stress, which is used in this criterion. The value of the material yield stress used is that calculated for the elevated temperature.

This criterion assumes elastic-perfectly plastic material where, at the yield stress, the pipe deformation increases infinitely with no increase of load. In reality, steels exhibit residual strength above the first yield, which is caused by material hardening.

The material ultimate tensile strength (UTS) is not used due to its uncertainty at elevated temperatures. This is due to a combination of the following effects:

• The uncertainty in the temperature-affected slope of the stress-strain curve up to the material yield;

• The uncertainty of the material yield stress value at elevated temperature; and

• The uncertainty of the temperature-dependent stress-strain curve from the material yield to the UTS value (i.e. the uncertainty of the temperature-dependent material hardening).



Temperature of blowdown pipe The temperature in the blowdown valve and pipe may rapidly drop to a sub-zero level due to the rapidly expanding stream of gas through blowdown valve. This needs to be checked and reflected in material selection because excessively low temperature may cause embrittlement of the valve and pipe material.

This requires an additional multi-physics analysis without fire loading and with PFP coating if required for fire protection.

Flange connections As described in Section 8.2.3 unprotected flanges affected by the impinging fire lose their tightness within a few minutes after the start of the fire. The integrity of a flange connection can be checked by the calculation of differential thermal expansion of flange components and strength of the nut thread [43 & 67]. The time to loss of tightness may be calculated based on lumped thermal mass and thermal expansion of the flange components and it should include:

• The heating-up of the flange body exposed to fire;

• The heating-up of the cylindrical parts of bolts shielded from the fire by the flange body;

• The differential thermal expansion of the flange components;

• Young’s modulus for the flange components varying with temperature;

• Pre-tension of the bolts at operating temperature;

• The increase of the bolt tension due to the differential thermal expansion of flange components; and

• The loss of strength of the thread between the bolt and the nut where the latter is exposed to impinging flame.

Valves and actuators Valves lose their tightness by differential thermal expansion of their components. Fire tests of valves are carried out to confirm the suitability of the valve for plant with potential fires. It should be noted that the fire tests may not represent the actual fire scenarios that the valve would be exposed to in case of a fire.

Similar to flanges, the time to loss of tightness of a valve may be calculated based on lumped thermal mass and thermal expansion. This method was proved to be adequate by comparing the results against valve tests [43].

Heat transfer calculations show that for an assembly of fire-protected valve with an unprotected actuator the rate of heat transfer from the actuator to the valve is such that the valve may lose its tightness within 15 minutes of the start of an impinging fire [43].

8.3.8 Mitigating options if acceptance not met

Various possible mitigating solutions If a rupture of piping occurs as a result of a combination of excessive heat load and internal pressure, an acceptance of the situation will have to be judged based on the risk analyses [65]. Residual quantities of hazardous fluids and escalation potentials both within the area and towards adjacent areas shall be taken into account.

Where rupture cannot be accepted, i.e. the risk acceptance criteria are not met, the provision of additional protective systems and arrangements shall be implemented. This can include one or more of the following options:

• The selection of materials with better performance at elevated temperatures;

• Choosing a pipe wall thickness that is greater than that required to meet process pressure design requirements. However, as a consequence the material may have lower fracture toughness at normal operating temperature and exhibit a higher propensity for brittle fracture;

• Increasing pipe wall thickness might delay failure, allow sequential depressurisation and result in reduction in peak relief rates with overall cost reduction;

• The selection of fire resistant valves might reduce the need for fire insulation on them with less severe consequences in case of failure. It should be checked, however, that the fire test of the fire resistant valve represents the actual fire load on the plant being designed;

• As unprotected flanges lose their tightness within a few minutes after the start of impinging fire, the use of fully welded pipe



instead of weld connections can reduce the risk of fire near a vessel. However, this needs to be weighted against the requirement for adequate means to isolate sections of pipework for maintenance. Careful location of flange connections can minimise the chance of any ignited leaks from these impinging on adjacent plant and equipment;

• Fire protection coating or cladding (fire insulation) will provide thermal protection to restrict the rate at which heat is transmitted to the object or area being protected. For the design of pipework against fire, fire insulation is used to:

Prevent escalation of the fire due to progressive releases of inventory from, e.g. loss of tightness of flanges;

Protect essential safety systems and components, such as shutdown valves, flanges and pipework; and

Minimise deflections by protecting the critical structural members and pipe supports.

• It should be noted that fire insulation has detrimental effects on explosion overpressure (by increasing pipe diameter), corrosion, the duration of plant maintenance and inspection activities, etc; and

• Modifications to the general arrangements and layout that have an impact on the time to rupture. This includes, but is not limited to the location and rating of firewalls and plated decks.

Pressure relief and depressurisation systems Pressure relief / depressurisation (blowdown) systems are normally protected from over-pressure by pressure relief systems such as relief valves and bursting discs, to automatically prevent overpressure during process operations and upset conditions. Depressurisation (blowdown) is the main method of preventing catastrophic pressure system failure offshore by intentionally reducing the inventory of the system in an emergency. Some onshore petrochemical plants also use depressurisation systems. Depressurisation reduces the system’s pressure thus decreasing the applied material stress, which can minimise or eliminate the risk of rupture – a benefit which is not normally found when relief systems are applied on their

own. It also reduces the inventory, which should decrease the duration of any consequent fire in the event of loss of integrity. Depressurisation can also be used during normal shut-down, for other upset conditions, or routine plant maintenance.

Relief and depressurisation systems lead gas or vapour to a vent or flare and these systems need to be designed using the multi-physics method including the pipework and flare knock-out drum. An increased size of relief or depressurisation orifice will enable a faster reduction of pressure, reduce the duration of fire and may prevent pipe rupture as the applied stress may be lower than the material yield stress throughout the required time for escape of personnel. The size of the flare header, the height and location of the flare stack and the capacity of flare tip need to be checked to confirm their adequacy for the increased maximum simultaneous relief / depressurisation rates.

Sequential relief / depressurisation may also be considered to keep relief / depressurisation rate within a pre-determined flaring capacity. The process segment the fire originates from and high risk process segments should be pressure relieved / depressurised first.

Depressurisation systems may be activated manually or automatically, where the latter is activated on the detection of fire. The reliability of fire detection devices need to be taken into consideration to avoid the activation of depressurisation system on spurious detection.

The vent / flare system must survive and remain functional throughout explosions and fires until the escape and rescue of injured personnel is complete and the personnel are in location of relative safety.

Prevention of additional releases The elimination of any additional releases resulting from an initial fire may require the fire protection of all pipework and flanges, which may be uneconomical in some cases. This would require a careful consideration of potential consequences. If affected by fire, flanges may lose their tightness and pipework may rupture, but it is unlikely that projectiles would form. Such escalated fire events may be controlled and limited with the fire area.



On the other hand projectiles may form as a result of pressure vessel rupture, where the projectiles may penetrate firewalls or damage equipment in neighbouring unit, and therewith escalating across the boundary of the fire area. Such projectile-generating events must be prevented.

8.4 Firewater and other essential safety systems

8.4.1 Firewater systems The strategy for protecting firewater systems against fires is to adapt the general procedures described in Section 8.3 as follows:

• Select the routing and location of the firewater supply pipework and deluge valves such that, for firewater supply to a fire area they are located in neighbouring fire area and normally segregated by a firewall or by sufficient distancing;

• The firewater system up to the deluge valves is normally pressurised whilst the pipework matrix leading to deluge nozzles and the nozzles themselves are dry; and

• Whether or not to fire-insulate the pipework matrix downstream the deluge valves is a matter of judgement. Firewater systems in offshore installations are normally activated automatically on the detection of fire in the area and water comes out of the deluge nozzles within a minute after the fire was detected. It is therefore considered that this time period is short enough to prevent the temperature rise of the deluge pipework that would render it inoperative. Also, the heat flux from the fire that the pipework matrix to nozzles will be exposed to will not be high, as it is unlikely that the pipework matrix will be impinged by the fire as a whole. Once firewater flows in the pipework matrix it provides adequate cooling to keep the water supply to the nozzles operative. Following from these considerations the pipework downstream deluge valves would not require fire insulation.

The materials of construction for firewater pipework normally are:

• Galvanised carbon steel; • Stainless steel; • Titanium alloy; and

• GRP.

The use of GRP pipes should be supported by fire tests as some GRP resins are flammable and their use should be avoided. Specialist advice should be sought prior to using GRP pipes.

8.4.2 Other essential safety systems In general, all essential safety systems should survive and remain functional throughout the explosion and fire scenarios they should prevent, control or mitigate. Such equipment include essential cables and instrumentation that may have much lower tolerance for temperature rise than a pipe and consequently will have a need for increased fire protection.

8.5 Pipe supports On fixed offshore installations pipe supports are normally connected to primary or secondary structural steel. On floating production units and onshore facilities pipes are supported by pipe racks.

Pipe supports play a key role in ensuring the integrity of the piping systems. Excessive deformations of the support structures and pipe racks must be prevented for pipework that is required to maintain its tightness when exposed to fire.

Guidance [9] on the protection of pressurised systems recommends that pipe / equipment supports and the secondary steel supporting these supports must maintain their integrity until such time when possible damage of the equipment / pipe they support is acceptable. For this reason, the supports have to be protected by fire protection coatings unless adequate integrity can be documented by analyses. If for any reason it is desired not to use PFP on all the pipe supports, it must be documented that the pipe integrity will be maintained without the presence of PFP on all pipe supports.

8.6 Flanges, bolts and welds 8.6.1 Flanges Flanges that are to maintain their integrity and resistance against leaks in fire must be protected by fire insulation.



8.6.2 Bolts and welds The recommendations for welds and bolts are drawn from [70] which is based on BS5950 Part 8 [71].

• It is generally accepted that welds behave in a manner similar to the parent material in fires. However, there is some conflicting evidence that suggests a significant reduction in weld strength after a fire. The failure mode associated with the fire is not known, neither is the actual condition of the weld before the fire;

• Bolts do not behave very well in fires, and the higher the bolt specification, the poorer the fire enduring qualities. The loss of strength of Grade 4.6 bolts follows that of Grade 43 steel. For Grade 8.8 bolts, the strength reduces after exposure to temperatures above 450ºC, to 80% at 600ºC and 60% at 800ºC;

• High strength friction grip bolts behave in a similar manner to Grade 8.8 bolts; and

• Higher specification bolts such as ‘L7’ and those formed from ‘Macalloy’ bars, etc., should be replaced as a matter of routine if fire damage has occurred. Alternatively, advice should be sought from the material suppliers. Note that nickel-based alloys may maintain good mechanical properties both during and after a fire.

Materials for flange bolting may be different from the above, and will usually be selected for compatibility with the specified piping material. For design, equivalent material data will have to be obtained. A typical bolt material for carbon steel piping is CR-0.2 Mo (ASTM A193, A194).

The Statoil / Norsk Hydro / Scandpower guidance [9] provides the following additional information on the behaviour of bolts:

• Bolted connections, including flanges and valve connections, must be verified with respect to the need for fire protection; and

• The temperature of the bolts must be kept below 500ºC. Unless bolted connections are specifically designed to withstand higher temperatures, the connections should be equipped with fire insulation coating.

8.7 Fire protection 8.7.1 Passive fire protection coatings This Section provides a brief discussion on passive fire protection while the following documents can be used for obtaining more detailed information on PFP performance.

• ISO standard 13702 (1999)[26]; • OTO 2000 051 [8], which provides a very

good background discussion on the use of PFP materials;

• Passive fire protection: Performance requirements and test methods (Appendix A, OTI 92 606 [72]) which appraises the performance requirements for offshore PFP systems and assesses the adequacy of the then current tests for ensuring that performance;

• Availability and properties of passive and active fire protection systems (Appendix A, OTI 92 607 [73]) which reviews a selection of the various types of passive fire protection products which are used on offshore structures. Appendix C of OTI 92 607 [73] contains a listing of manufacturers, products and product properties;

• The Interim Guidance Notes [2] give an indication of how the information given in these reports should be applied; and

• Chamberlain [74] provides a brief discussion on the recent findings from various JIPs related to the development of jetfire testing procedures for passive fire protection materials.

Background Passive fire protection is defined, in the guidance [26], as “a coating, cladding or free-standing system which, in the event of a fire, will provide thermal protection to restrict the rate at which heat is transmitted to the object or area being protected”. These materials are used to:

• prevent escalation of the fire due to progressive releases of inventory;

• protect essential safety systems and critical components such as separators, risers and topside emergency shutdown valves and their actuators;

• minimise damage by protecting the critical structural members, particularly those



which support the temporary refuge, escape routes and critical equipment; and

• protect personnel until safe evacuation can take place.

The required fire resistance may be achieved by the use of PFP in conjunction with active fire protection systems such as water deluge, in which case a minimal residual protection must be achieved should the active systems fail to operate. PFP is used particularly where active systems are impracticable, have insufficient reliability or where protection is needed within the probable response time of an active system.

Types of PFP There are many types of PFP materials on the market, which can be broadly categorised into groups as follows [8]:

• Spray-applied and coating materials; • Blanket / flexible jacket / wrap around

systems; • Prefabricated sections; and • Enclosures and casings.

Functional requirements In ISO 13702 (1999) [26], the following functional requirements are given:

• PFP shall be provided in accordance with the Fire and Explosion Strategy (FES);

• PFP of essential systems and equipment, or enclosures containing such systems and equipment, shall be provided where failure in a fire is intolerable;

• where PFP is required to provide protection following an explosion, it shall be designed and installed such that deformation of the substrate caused by an explosion will not affect its performance;

• selection of the PFP systems shall take into account the duration of protection required, the type and size of fire which may be experienced, the limiting temperature for the structure/equipment to be protected, the environment, application and maintenance; and

• smoke and fume generation in fire situations.

PFP materials should be approved for their intended use. Various "approved lists" (see DNV

[75] and LR [76]) exist which contain general data such as name and location of manufacturer, brief description of product, areas of application and type of certification. Where general approvals from a recognised third party or governmental body are not available, PFP fire performance should be documented by test reports from a recognised fire test laboratory. Jet fire resistance tests have been developed for this purpose as discussed in the following subsection.

Fire resistance tests OTO 2000 051 [8] provides a very good review on test for passive fire protection material. It identified the development of the Jet Fire Resistance Test of Passive Fire Protection Materials (JFRT, OTI 95 634 [77]) as a key improvement. This test involves use of a sonic, vapour-only 0.3 kg/s propane jet fire. The test was shown (OTO 97 079 [78]) to reproduce key conditions typical of large scale fires resulting from high pressure releases of natural gas and is now widely used to assess PFP coatings and systems.

More recent work [79] has been carried out by the Health and Safety Laboratory which proposed a number of changes from the procedure originally developed and published by the HSE in [77]. These changes have been incorporated into a draft British Standard version of a test procedure for the determination of the resistance to jet fires of passive fire protection materials [80]. Following Roberts [79], the four versions of the original test, intended primarily for coating systems, are:

• Panel test that applies to cases involving panel material used to form the rear wall of the flame circulation chamber;

• Planar steelwork test used for PFP material applied to steelwork with no corners and edge features and to cylindrical vessels, pipes and tubular sections of outside diameter greater than 1.0m, and hence where the surface may be considered as planar;

• Structural steelwork test used for PFP material applied to steelwork with corners or edge features such as I beams; and

• Tubular section test used for PFP materials applied to cylindrical vessels, pipes and tubular sections of up to 0.50m outside diameter.



A key feature of the draft standard [80], as reported by [79], is the extension of the original procedure to cover the assessment of passive fire protection systems in an assembly test.

8.7.2 Determination of the required PFP performance

Determination of PFP type and thickness The normal procedure is for the process system and piping designer to specify the fire type and intensity, and the required endurance together with the maximum allowable temperature of the pipe material in the dimensioning fire scenario. The designer is also expected to specify the coatback requirements (length of continuation of the PFP application beyond the heat protected surface).

The maximum allowable temperature is obtained from the multi-physics analysis or the section (Hp/A) method as described in Section 8.3.6.

The PFP supplier will normally select the system type, its thickness and the design of the casings (cast PFP items) to piping equipment that is required to meet the specified criteria.

Preliminary information on the capability of a material to provide thermal insulation may be obtained from thermal diffusivity, which is defined as:

pCk

ρα =

where

α is the thermal diffusivity (m2/s);

k is the thermal conductivity (W/mK or J/msK);

ρ is the density (kg/m3); and

Cp is the specific heat (J/kgK).

The physical significance of thermal diffusivity is associated with the propagation of heat into the material during changes of temperature with time. The higher the thermal diffusivity, the faster the propagation of heat into the material.

It should be noted that the thermal diffusivity does not normally contain the thickness of the insulation material or the variations of specific heat, density and thermal conductivity with temperature. The thickness of the insulation

material may also change when affected by elevated temperature and the final decision on what material should be used and its thickness should be left to the PFP supplier, who will also provide the performance warranty.

Additional factors The draft British Standard [80], as reported by Roberts [79], includes a new section giving advice on additional factors to be considered when assessing performance by test. For coating and spray materials the following factors are considered:

• Substrate temperature, where the location and time of any sudden increase in the rate of temperature rise, is indicative of failure of PFP coating at that point;

• Reacted / un-reacted remaining material and condition of reinforcement; and

• For systems and assemblies, the corresponding considerations provided in [79] include:

Substrate temperature; and Loss of integrity.

Design and performance requirements ISO 13702 (1999) [26] includes an appendix which gives typical fire integrity requirements. For example, for load bearing structures in process areas, resistance to a one hour jet fire at a critical temperature of 400°C is required. The jet fire resistance test (JFRT) is mentioned as a suitable test. The reference temperature of 400°C was used as a typical value for structural steel. For aluminum, the corresponding temperature is 200°C and, for other materials, the critical temperature is the temperature at which the yield stress is reduced to the minimum allowable strength under operating load conditions.

Few recent articles have tried to address the issue of a risk-based Design Approach for passive fire protection (e.g. Yasseri, FABIG Article No. 2001 [81]).

Weathering of PFP Roberts [82] presented results of tests on PFP subjected to weathering and corrosion in the FABIG Technical Meeting on Passive Fire Protection. Initial conclusion included:

• Jet fire resistance test:



1. Limited temperature variation between weathered and new specimens;

2. Little difference in Char formation between weathered and new; and

3. Large difference in Char formation at top and bottom positions.

• Corrosion: 1. All specimens showed some corrosion

from edges; 2. Some specimens heavily corroded; and 3. Method of application is critical.

The presence of corrosion means that PFP may lead to increased leak frequency. The risk of corrosion requires additional inspection, whereby the PFP coating needs to be removed and re-applied. This leads to increased number of personnel in hazardous areas and the increase of platform and occupational risks.

Concerns related to use of PFP The main concerns relating to use of PFP include [9]: • Increased corrosion of materials covered by

PFP; • Performance of weathered PFP; • Reduced possibilities for inspection and

maintenance of equipment covered with PFP;

• Increased congestion and blockage for explosion overpressure;

• Increased weight; • Increased take-up of space; • Increased need for maintenance of the PFP;

and • Increased cost.

Specific concerns related to PFP on piping While most guidelines for equipments are related to protection of vessels, it should be recognised that vessels and piping will behave differently in a fire situation due to different surface area to volume ratio [9]. The consequences of rupture would also be different for vessels and pipes. For pipes there is obviously a difference between gas filled and liquid filled lines. Due consideration should in this context be paid of “self draining” pipes, i.e. pipes that normally are completely or partially

liquid filled could be dry in a shut down situation [9].

For pipe branches the potential problem of heat conducted from the branch pipe into the main pipe, which may cause a hot spot and material weakening in the main pipe, should be considered. A PFP coatback of the branch pipe up to the first flange may suffice, however, the implications of the unprotected flange losing its tightness have to be taken into consideration.

8.7.3 Active fire protection Background The FABIG Technical Meeting on Mitigation [83] identifies the following main categories of active fire protection:

• Water deluge (general area, equipment specific, curtain, and hybrid);

• Foam systems; and • Fire monitors for manual fire fighting.

The primary form of fire protection to processing areas is water deluge by water spray, where fixed deluge systems are provided to:

• Control pool fires and thus reduce likelihood of escalation;

• Provide cooling of equipment not impinged by jet fires;

• Provide a means to apply foam to extinguish hydrocarbon pool fires; and

• Limit effects of fire to facilitate emergency evacuation, escape and rescue operations.

The Fire and Blast Engineering Project Phase II [3] investigated the effect of water deluge on confined pool and jet fires. The main findings were:

• The well ventilated jet fires were not extinguished by typical offshore water deluge. The jet fires continued to burn at the same rate but there was a substantial reduction in fire intensity (heat flux);

• Fuel controlled (under-ventilated) jet fires were controlled but were not extinguished when deluge was activated soon after ignition;

• Fuel controlled (under ventilated) jet fires were extinguished when deluge was activated 10 to 12 minutes after ignition and the fire compartment was hot;



• There was no significant difference between the effects of water deluge on vertical and horizontal jet releases;

• It is possible for the fire to re-ignite after the water deluge is terminated due to the presence of hot gases and surfaces in contact with fuel;

• Extinguished jet fires represent a potential explosion hazard if the fuel continues to be released; and

• Generally confined pool fires are not extinguished by water deluge, but the fire is controlled and burns at a much lower rate.

OTO 2000 051 [8] identifies four broad types of deluge systems:

• Area protection designed to provide non-specific coverage of pipework and equipment within process areas;

• Equipment protection designed to provide dedicated coverage of critical equipment such as vessels and wellheads;

• Structural protection designed to provide dedicated coverage of structural members; and

• Water curtains to reduce thermal radiation.

In a FABIG newsletter article, Shirvill and Lowesmith [84] reported on a major JIP study. The main factors studied in the first phase include:

• The effect of water deluge coverage rate; • The effect of the size of the pool fires; • The effect of weather conditions; and • An assessment of the differences between

mitigating effects of sea and fresh water.

The work showed that substantial benefit may be gained particularly in reducing the thermal radiation field around a fire. In the case of pool fires the deluge was also shown to reduce the size of the fire and, in certain circumstances, the interaction of the water with the pool lead to extinguishment. A further benefit was in the reduction in smoke levels within and beyond the test rig.

The issues studied in the second phase of the work include:

• Stability of gas jet fires in the presence of deluge;

• The effectiveness of deluge on condensate pool fire and crude oil jet fire;

• The effectiveness of the spray generated by various nozzle types; and

• Vessels and pipe targets were included in the fires to allow an assessment of the ability of both area and dedicated deluge to provide protection to objects engulfed by pool or jet fires.

Design guidance The Phase I Fire and Blast Joint Industry Project [1] refers to Department of Energy Guidance ‘Offshore Installations: Guidance on Fire Fighting Equipment’ Note SI 611. In this guidance a general water application rate of 12.2litres/min/m2 is recommended.

The Interim Guidance Notes [2] quotes the following additional rates:

• 10 litres/min/m2 to protect against pool fires;

• 20 litres/min/m2 to protect against high pressure jet fires;

• 400 litres/min/m2 to protect against high pressure jet fires impinging on structural steelwork and vessels; and

• 400 litres/min to each wellhead.

OTO 2000 051 [8] presents a brief discussion on the application of water sprays for specific deluge on equipment. It is stated that the water spray design should surround the equipment with medium velocity nozzles spaced at 2.0 to 2.5m intervals and 0.6m from the surface. Complex-shaped objects would be covered by directing the spray at a virtual box enclosing the object under consideration.

The offshore industry often uses firewater rates recommended by the American National Fire Protection Association (NFPA) [85].

Concerns related to use of active fire protection In the FABIG Technical Meeting on Mitigation, Renwick [83] identified various concern areas in the application of active fire protection. Most of these concerns are related to verification issues such as:



• Reliability of water supply; • Time to full activation; • Nozzle blockage; • Blast resistance; • Fire damage to dry pipework; and • Damage tolerance.

In addition, the following special cases requires more attention:

• Impinging jet fire: deluge likely to be ineffective;

• Partially confined jet fires: deluge may be positively dangerous (extinguish flame and increase explosion risk); and

• Wellbay / Xmas tree fires: very high water application rates often specified but purpose unclear.

8.8 Optimisation of fire protection

Based on the unwanted effects of PFP coatings, limitations that may exists in the flare capacity and costs, projects are interested to use minimum PFP coatings and small blowdown orifices providing that required safety levels are maintained.

One of the aims of the IP [10] and Statoil/Norsk Hydro and Scandpower [9] guidelines is to minimise PFP coatings to alleviate the problems associated with them.

Optimum PFP coatings (or no coatings at all) and pressure relief / blowdown orifices for pressure vessels and pipework can be designed

using iteratively the procedure illustrated in Figure 8.5.

Safety barriers should work together in a harmonised manner to resist an accidental fire, and it should be borne in mind that a minimum pressure relief blowdown orifice and a minimum thickness of PFP coating are just two of many other safety barriers. The actual performance of unprotected flanges, pipe supports, reliability of the firewater system (if the latter is taken a credit for), etc. have all to be taken into account.

The paper by Medonos and Geddes [86] stresses the importance of using verified methods in the design optimisation.

Once the analytical work is nearing completion, the material specification and selection is conducted with overlap back to the analysis process. At this stage the project / facility specific requirements need to be taken into account. The selection process flow chart in Figure 8.6 shows the steps that require to be taken to achieve an optimised PFP solution.

Figure 8.7 shows many of the elements that have an influence on the cost contribution of PFP to the project. As much as 80% of the materials and installation cost are frozen and cannot be changed by the time the project gets to the procurement phase. It is therefore important to understand the impact of design decisions in the total cost. The ongoing maintenance and life cycle costs are often overlooked at the procurement stage.



Figure 8.6 Material selection process for fire protection coatings



Figure 8.7 Input to and cost of the selection of fire protection coatings



9. DESIGN OF PIPE SUPPORTS AGAINST FIRES AND EXPLOSIONS

9.1 Introduction Pipe supports may be divided into the following broad components:

• Welds and bolts that attach the piping systems to support arrangements;

• The steel components that provide direct support to the pipes;

• The secondary steelwork that supports the direct support components; and

• Bolted connections of the pipework systems.

Pipe supports are used to support the weight of piping runs, the associated valves and the contained fluids. Pressurised piping systems that contain fluid, particularly gas, are likely to have a low weight per unit length of pipe compared to the pipe self-weight.

The pipe supports play a key role in ensuring the integrity of the piping system. Flanges may leak or pipework may rupture if subjected to large strains, which could develop if one or a number of pipe supports were to fail. It is normal practice for piping containing hydrocarbons in hazardous areas to be joined by welding wherever possible, partly to provide greater ductility should supports fail.

In some cases the pipe supports have a multi-purpose function and support a number of different services. For example, cable ladders and HVAC ductwork may be supported along with piping.

The recent trend has been to use “multi-discipline” supports for major pipe racks, which will also support electrical and instrument cables in addition to pipes. This concentration of services may present a greater hazard in a local fire, but should be easier to provide total protection.

Section 9.2 presents an overview of some of the support types used on offshore platform. Section 9.3 provides guidance on ductile construction, specific to the types of supports discussed in Section 9.2.

9.2 Types of pipe supports In general, pipe supports are of base plate or hanger type, both of which provide guide and anchor support. Typical details are:

• Trunnion Base Plate with support and slide guide units (Figure 9.1);

• Trunnion Base Plate with support and stop slide units (Figure 9.2);

• Trunnion Base Plate with support, guide and stop slide units (Figure 9.3);

• Adjustable Trunnion Base Plate (Bottom Plate only) (Figure 9.4);

• 4-Bolt clamps for Copper-Nickel Lines 8 Inches and above (Figure 9.5);

• 3-Bolt Clamps for stainless steel, Duplex, Galvanized and Acoustic Insulated lines (Figure 9.6);

• Lateral support without vertical restraint (Figure 9.7); and

• Fixed support configuration (Figure 9.8).



Figure 9.1 Type 1: Trunnion base plate with support and slide guide units

Figure 9.2 Type 2: Trunnion base plate with support and stop slide units



Figure 9.3 Type 3: Trunnion base plate with support, guide and stop slide units

Figure 9.4 Type 4: Adjustable Trunnion base plate (bBottom plate only)



Figure 9.5 Type 5: 4 bolt clamps for Copper-Nickel lines 8 Inches and above

Figure 9.6 Type 6: 3 bolt clamps for Stainless Steel, Duplex, Galvanised and acoustic insulated lines



Figure 9.7 Type 7: Lateral support without vertical restraint

Pipe sleeve

Pipe

Bolt

Stiffener

Pipe

Pipe sleeve

Base plate

Figure 9.8 Fixed support configuration

9.3 Guidance for ductile construction

Section 6 provided general recommendation for ductile construction. This Section briefly discusses additional guidance on ductile construction for the types of supports discussed in Section 9.2 above.

Both base plate and hanger type supports consist of bolts and steel components (plates or rods). In both cases, to ensure a ductile failure mode, the capacity of the bolts should be greater than the plastic moment of the attached plate or rod. The plastic capacity of the plate or rod should be an upper bound value taking account of yield variation and strain rates.

Particular care should be taken to ensure that the steel connections could withstand both the dynamic loads and any load reversals imposed on them. The parameters that may be varied to ensure that plate / rod failure will occur before bolt failure are:

• Bolt spacing • Thickness of plate / rod • Shape of stress strain curve of material



10. TYPES OF PIPING AND MATERIAL PROPERTIES

The main types of steel used offshore are identified in this section. The relevant Codes and Standards are reviewed, and the high strain rate properties of commonly used offshore steels are presented. A summary of piping material that are commonly used offshore, and of the material properties that are required for carrying out a blast assessment, is presented in this Section, while a more detailed discussion is provided in FABIG Technical Note 6 [7].

10.1 Standards used for piping and piping material

There are many different standards for piping and piping materials depending on the functionality of the piping. The main standards for steel piping for carrying combustible fluids are as follows:

• API Specification 5L Specification for Line Pipe [87]:

The purpose of this standard is to provide standards for pipe suitable for conveying gas, oil and water in both the oil and natural gas industries. It covers seamless and welded steel line pipes. It also covers a variety of steel grades including X42 through to X80;

• BS EN 10208: Flat products made of steels for pressure purposes – Part 2:1993 Non-alloy and Alloy steels with specified elevated temperature properties [88, 89 and 90]:

The purpose of this code is to specify requirements for flat products for pressure purposes made of weldable non-alloy and alloy steels with elevated temperature properties. It covers a variety of steel grades including P235GH to P355GH;

• ASTM A106: Standard specification for seamless Carbon steel pipe for high-temperature Service [91]:

This specification covers seamless carbon steel pipe for high-temperature service for Steel A106 Grade A, A106 Grade B and A106 Grade C;

• ASTM A333: Standard specification for seamless and welded steel pipe for low temperature service [92]:

This specification covers seamless and welded carbon and alloy steel pipe intended

for use at low temperatures. It covers a variety of Grades from A333 Grade a through to A333 Grade 11;

• ASME B31.3 Process piping – Chapter III materials [11];

• BPVC Section III – Rules for the construction of nuclear power plant components – Div 1 – Subsection NC-Class 2 components and subsection NC – Class 3 components [93]:

This code provides a useful example of how a risk based methodology may be used for the analysis of piping against fire and blast;

• Structural analysis and design of nuclear plant facilities, 1980, Committee on Nuclear Structures and Materials of the Structural Division, ASCE [94]:

This code provides a useful example of how a risk based methodology may be used for the analysis of piping against fire and blast; and

• FABIG Technical Note 6 [7]: Elevated temperature and high strain rate material property data of offshore steels:

The purpose of this document is to provide guidance on elevated temperature and high strain rate property data that is currently available for high strength steels used specifically for offshore structures. The document covers a variety of steel grades including Grades 355EMZ and 450EMZ.

For piping that does not convey combustible fluids, the following product standards may be used:

• BS EN 10216-1:2002 and BS EN 10217-1:2002 Welded and seamless steel tubes for pressure purposes [95 and 96];

• BS EN 10216-2:2002 and BS EN 10217-2:2002 Welded and seamless steel tubes for pressure purposes with specified elevated temperature properties [97 and 98];

• BS EN 10216-4:2002 and BS EN 10217-4:2002 Welded and seamless steel tubes for pressure purposes with specified low temperature properties [99 and 100];

• BS 3604 Steel pipes and tubes for pressure purposes. Ferritic alloy steel with specified elevated temperature properties [101]; and



• BS 3605 Austenitic stainless steel pipes and tubes for pressure purposes [102].

10.2 Typical materials used The following material grades are commonly used offshore:

• API 5L316L; • ASME B31.3X52/X65; • API 5L Carbon Steel A333 Grade 3 and

Grade 6; • Duplex Stainless Steel; • GRP (24 and 18 inches); and • Copper nickel for water deluge.

Table 10.1 summarises the typical material for piping used offshore and relates it to the relevant standard discussed in Section 6.1 above.

10.3 Required material properties for carrying out blast assessment

10.3.1 Strength enhancement factors

For general design, the design strength σy is taken as either the yield strength or, the minimum specified proof strength. However, when blast loading is being considered, the design strength may be enhanced to σdyn, to take advantage of the improvement in strength due to the high strain rates. Thus:

( )SRydyn Kσσ = for carbon steels; and

( ) 20.SRydyn Kσσ = for stainless steels.

Where KSR is the strength enhancement factor at yield, or corresponding to a particular strain.

The enhancement of stresses as a result of high strain rates can also be represented by the Cowper - Symonds empirical relationship. The Cowper-Symonds constants D and q, for 316L, SAF2304 and 2205 stainless steels, which have been obtained from a least mean squares fit, are given in Table 10.2.

Material properties at high strain rates are required for determining the response of piping to explosions. The Table below provides a

summary of available high strain rate material properties data for the commonly used offshore steels. It can be seen that there is sufficient data for stainless steels 316L and duplex as a result of work carried out by the SCI and reported in FABIG Technical Note 6 [7]. Further research and experimental tests may be required to generate the missing data, marked with a x , in Table 10.3.

10.4 Required material properties for carrying out fire assessment

There are no elevated temperature material properties data for the steel pipes used for carrying combustible fluids. For piping that does not convey combustible fluids, elevated temperature material property data are available. However, they are based on isothermal or steady state test methods. In BS EN 10216-2 [97], BS EN 10217-2 [98] and BS 3604 [101], the data are minimum guaranteed 0.2% proof strength, whilst in BS 3605 [102], the data are minimum guaranteed values 1.0% proof strength. FABIG Technical Note 6 [7] provides data on elevated temperature properties for high strength steels used offshore and discusses two types of tests for obtaining these properties. A summary of this data, relevant to offshore piping is presented below.

10.4.1 Summary of available data Table 10.4 below provides a summary of available material mechanical properties at elevated temperatures for the commonly used offshore steels. Further research and experimental tests may be required to generate the missing Figure 2.3 data, marked with a x, in Table 10.4.

In the absence of available data on some of the steel types below, it is suggested that data for steels with similar compositions be used (where possible).

• For grade X52 use material data for Grade 355;

• For grade X65 use material data for Grade 450; and

• For Grade A333 Grade 6 use material data for equivalent grades in EC3.



Table 10.1 Typical material used for piping

Code Grade API 5L BS EN 10208 ASME B31.3 BPVC Section 3 FABIG Technical Note 6

316L A x x x A X52/X65 x x A x x A333 Grades 3 and 6 A x x x x Duplex Steel x x x x A GRP x x x x x Copper nickle x x x x x

Table 10.2 Cowper - Symonds constants for stainless steels

Material Proof strength

D s-1 q σo

MPa 316L 0.1% 471 5.76 263

0.2% 240 4.74 277 SAF 2304 0.1% 22.0 2.51 516 635 (alt) 4.04 (alt) 0.2% 3489 5.77 527 2205 (318) 0.1% 769 5.13 544

0.2% 5958 6.36 575

Cowper Symonds relationship: q

s

dyn

D

1

1 ⎟⎠⎞

⎜⎝⎛+=

εσ

σ &

Where dynσ is the dynamic stress at a particular strain rate, ε& and

sσ is the static stress Table 10.3 Availability of high strain rate material property data

Steel Property

316L X52/X65 (API 5L)

A333 Grades 3 & 6

Duplex GRP Copper-Nickel

Elasticity Modulus A x x A x x Poisson’s Ratioi x x x x x x UYS A x x A x x LYS A x x A x x UTS A x x A x x Rupture Strain A x x A x x Stress-Strain curves A x x A x x Notes: i. It is assumed that Poisson ratio does not vary with strain rate

Table 10.4 Availability of elevated temperature material property data

Steel Property

316L X52/X65 (API 5L) A333 Grades 3, 6 Duplex GRP Copper-Nickle

Modulus of Elasticity, Eθi A x x A x x

Poisons ratio, νθ A x x A x x UYS, σyθ A x x A x x ULS A x x A x x UTS A x x A x x Rupture Strain, εu A x x A x x Stress-Strain curves A x x A x x Notes: i. It is assumed that Poisson’s ratio remain constant with temperature variation



11. OUTSTANDING ISSUES

The guidance in this Technical Note does not cover all issues relevant to the design of piping systems for fire and explosion loading. The outstanding issues are given in Sections 11.1 and 11.2 for explosions and fire, respectively. Section 11.3 addresses outstanding issues for piping supports, while Section 11.4 addresses outstanding issues for material properties at high strain rates and / or elevated temperatures.

11.1 Explosions The following outstanding issues should be addressed:

• Develop a methodology for a simple and accurate way for choosing flanges that are stronger than pipes, to ensure that pipe failure will occur before flange failure. This will contribute towards achieving a ductile behaviour of the piping system. This may be achieved by:

Developing a series of nomographs of bending strength of API flanges at different pressure; and

Developing a series of nomographs of piping yielding strength at different pressure.

• Assess the accuracy of using ‘conservative’ drag coefficients to account for both drag and overpressure terms of the loading due to the shock wave;

• Develop a better way of calculating drag coefficients for individual pipes and methods for calculating explosion loads on pipe groups. Develop and test micro control volume formulation option in existing CFD programmes to allow pipes to be modelled explicitly in CFD (main or sub-models) rather than as sub-grid elements and perform explosion tests to calibrate for a wide range of explosion situations. Include some large scale tests to validate the whole process and provide public domain validation data; and

• Develop a methodology for treating pipe groups. Drag coefficients values for closely spaced piping systems are required. Also required is a method for determining the cut-off value when it becomes necessary to

include the piping system as one group rather than as individual pipes.

11.2 Fires The following outstanding issues should be addressed:

• Tabulated data on fire intensity within the flame and within the flame footprint onto a target are available with adequate accuracy. A flame impacting on a target spreads and the heat flux applied on the target in the area immediately outside the flame footprint exhibits rapidly tailing-off values. Only a limited information is available for this area;

• Little information exists on the performance of pressure relief devices under fire engulfment conditions. Such tests may need to be developed to ensure that the devices will operate in a satisfactory manner under fire loading;

• Standard fire tests exist for valves, but they do not cover the high fire intensity that may be experienced with jet fires. Such tests may need to be developed;

• Ambiguity remains on which method of assessing allowable stress is the most appropriate in response to fires and this topic requires further work. The influence of stress risers (such as: 1. constraining pipe supports, pipe branches, etc., 2. flanges, and 3. vapour-liquid interface) should be addressed in respect to their criticality in assessing the failure [10];

• At present, there are insufficient data to allow the designer to assess the adequacy of water deluge protection of pipework against a jet fire. There is a requirement for improved design capability of deluge systems so that a minimum required amount of water flowing over the surface of the pipe can be specified and delivered. These minimum amounts need to be linked with the fire scenario to be protected against and validated [10];

• Little is known about the relationship between passive and active fire protection. Area deluge is in common use offshore. Water monitors and fixed spray heads are



used and typically involve using water streams delivered at 10barg pressure, which could seriously affect the performance of epoxy intumescent and subliming PFP coatings. The effects of directed water deluge on passive fire protection under fire conditions require study. This is especially important where there is a desire to achieve the required level of fire protection by a combination of PFP and AFP systems [10];

• The ability of a fire protection medium to remain essentially in place after an explosion is fundamental to its potential for providing protection. A variety of test methods have been used to simulate blast effects ranging from mechanical deformation of a substrate using a hydraulic ram to gas explosions on full-scale panels. However, there are no agreed scenarios to test against (1.5barg has been used in the past) and hence there is no standardised, validated test method available [10];

• There is a concern as to how plant and equipment protected by PFP can be adequately inspected for corrosion and surface crack growth. Although there are a number of potential NDT techniques available, it is currently unknown whether these can be effectively used with PFP in situ [10];

• Some, mainly onshore, pipework contain pressurised reactive chemicals. Involvement in a fire could initiate runaway reactions. Methods need to be developed and validated for determining the temperature at which a runaway could begin and for specifying, if possible, fire protection to prevent runaway or to account for it in the sizing of emergency depressurisation [10]; and

• Develop a PFP application guide to provide recommendations on the different types of materials, including method of application and treatment of interfaces.

Fire loading on a target depends on such phenomena as the level of ventilation, the spatial distribution of the flame and the size of the target. As such it would be beneficial to specify

fire loading on a probabilistic basis. The information currently available does not cover this, however, the probabilistic determination of fire loads is being developed in a Joint Industry Project managed by Statoil [103].

11.3 Supports The following outstanding issues should be addressed:

• Little information, relating to the elevated temperature properties of welds or bolts, is available. The yield-strength reduction factors for welds and bolts are given in BS 5950: Part 8 [BS5950]. Phase I reports [1] states that bolts do not behave very well in fires, and the higher the bolt specification, the poorer the fire enduring qualities; and

• There is a conflict in the literature about the way in which welds behave in fire conditions [1]. A study carried out by the Department of Energy examined some fire-damaged tubular elements and concluded that the fire had reduced the basic strength of the welds, whereas BS 5950-8 suggests no change in performance after the cooling down period, for the commonly used structural steels. Tests on welded joints at elevated temperatures should be carried out taking into account the typical processes and details in use for pipe construction.

11.4 Piping materials The following outstanding issues should be addressed:

• To develop high strain rate and elevated temperature properties for commonly used piping steels on both offshore installations and onshore plants;

• Low temperature data are required if excessive cooling occurs during emergency depressurisation; and

• High temperature data are required if the system is to be designed to withstand a significant fire loading.



REFERENCES

1 Blast and Fire Engineering for Topside Structures, Phase I, 1991, The Steel Construction

Institute. 2 Interim Guidance Notes for the Design and Protection of Topside Structures against Fires and

Explosions, The Steel Construction Institute, Publication Number SCI-P-112, 1992. 3 Blast and Fire Engineering Project for Topside Structures Phase II, The Steel Construction

Institute, 1998. 4 CMR Gas Explosion Handbook, Bergen, April, 1997. 5 Design of Offshore Facilities to Resist Gas Explosion Hazard – Engineering Handbook, Ed. J

Czujko, Norway, 2001. 6 Explosion Loading on Topsides Equipment, Part 1, OTO Report No. OTO 1999 046, The Health

and Safety Executive, 1999. 7 FABIG Technical Note 6: Material properties of offshore steels for fire and explosion resistant

design, The Steel Construction Institute, 2001. 8 Review of the Response of Pressurised Process Vessels and Equipment to Fire Attack, OTO

Report No. OTO 2000 051, Health and Safety Executive. 9 Guideline for the Protection of Pressurised Systems Exposed to Fire, Scandpower, Statoil and

Norsk-Hydro, 13 May 2002. 10 Guidelines for the Design and Protection of Pressure Systems to withstand Severe Fires, The

Institute of Petroleum, 2003. 11 ASME B31.3 Code Process Piping Section III Materials. 12 Explosion Pressure Evaluation in Early Project Phase, OTO Report No. OTO 1999 048, The

Health and Safety Executive, 1999. 13 Anticipating and Coping with growth of Explosion Pressures through the Project design Phase, B

Brewerton, Natabelle Technology Limited / CMR, Notes of a presentation at the FABIG Technical Meeting on Inherently Safe Design, Fire and Blast Information Group, The Steel Construction Institute, 1997.

14 Guidelines for fire and explosion hazard management, UK Offshore Operators Association,

UKOOA 1995. 15 Successful Health and Safety Management, HSE Books, HSG 65, Health and Safety Executive,

1997. 16 Prevention of fire and explosion, and emergency response on offshore installations; Offshore

Installations PFEER Regulations 1995, Approved Code of Practice and Guidance. 17 A Guide to the Offshore Installation Regulations 1992, HSE Publications, Health and Safety

Executive, Guidance on Regulations L30, 1992.



18 Reducing Risks, Protecting People, HSE’s decision-making process, HSE Books, Health and

Safety Executive; 2001. 19 Guidance on ALARP for Offshore Division Inspectors making an ALARP Demonstration

(SPC/ENFORCEMENT/38), Hazardous installations Directorate, Available from HSE website. 20 Hazard Management in Structural Integrity – Volume 1, Main Report, Offshore Technology

Report OTO 98 148, Health and Safety Executive, 1999. 21 Hazard Management in Structural Integrity – Volume 2, Summary Report, Offshore Technology

Report OTO 98 149, Health and Safety Executive, 1999. 22 Hazard Management in Structural Integrity – Volume 3, Hazard Management Measures, Offshore

Technology Report OTO 98 150, Health and Safety Executive, 1999. 23 Hazard Management in Structural Integrity – Volume 4, Inherent Safety, Offshore Technology

Report OTO 98 151, Health and Safety Executive, 1999. 24 Guidance on ‘As Low As Reasonably Practical’ (ALARP) Decisions in Control of Major

Accident Hazards (COMAH) – (SP/Permissioning/12), Health and Safety Executive, available from the HSE website.

25 Pre 1988 – Early work on Explosions, M Vasey, British Gas plc, Notes of a presentation at the

FABIG Technical Meeting on Explosions, Fire and Blast Information Group, The Steel Construction Institute, 1994.

26 Petroleum and Natural Gas Industries – Control and mitigation of fires and explosions on

offshore production installations – Requirements and Guidelines, ISO 13702, 1999. 27 Salzano E, Cozzani V, The Use of Probit Functions in the Quantitative Risk Assessment of

Domino Accidents Caused by Overpressures, Safety and Reliability ESREL 2003 Conference Proceedings, Volume 2, A.A. Balkema Publishers, ISBN 5809 596 7.

28 Mercx WPM, De uitwerking van explosie-effekten op constructies, Juni 1988. 29 Medonos S, Improvement of Rule Sets for Quantitative Risk Assessment in Various Industrial

Sectors, Safety and Reliability ESREL 2003 Conference Proceedings, Volume 2, A.A. Balkema Publishers, ISBN 5809 596 7.

30 Petroleum and natural gas industries – specific requirements for offshore structures – Part 3:

Topsides structures, ISO 19901-3, 2002. 31 The Control of Major Accident Hazards Regulations 1999, Statutory Instrument 1999 No. 743,

Health & Safety Executive. 32 Waldram I M, Kurcharski M Finucane M, Industry Guidance on Fire and Explosion Hazard

management, Session 1, Paper 2, Second International Conference on Offshore Structural Design against Extreme Loads , ERA Technology, UK, 1993.

33 UKOOA / HSE, Update Guidance for Fire and Explosion Hazards, MSL Engineering

Consortium, CTR 103, Explosion Hazard Philosophy. 34 FABIG Technical Meeting on Guidance, Codes and Regulations for the Design of Offshore

Structures Against Fires and Explosions, 1995, J Clegg, Fire and Explosion Aspects of UK PFEER Regulations, 1995.



35 FABIG Technical Meeting on Risk Based Design, 1999, The Development of Design Philosophy

Guidance for Fires and Explosions, D Stirling. 36 Area Classification Code for Petroleum Installations, Part 15 of the Institute of Petroleum Model

Code of Safe Practice in the Petroleum Industry. 37 "Regulations relating to health, environment and safety in the petroleum activities", Norwegian

Petroleum Directorate, entered in force 1st January 2004. 38 Explosion Hazard Management for the Shearwater Project, Savage and Harper, ERA Major

Hazards Offshore 1999. 39 FPSO Girassol, Tonda H, Explosion and Fire Analysis for FPSO Girassol, off Angola, Tonda H,

Notes of a presentation at the FABIG Technical Meeting on Fire & Blast Issues on Deepwater Platforms, Fire and Blast Information Group, The Steel Construction Institute, 2001.

40 Fire and Blast Information Group, FABIG Technical Note 2, Explosion Mitigation Systems,

1994. 41 Vector International, Products and Services for Industrial Piping Applications Onshore, Offshore

and Subsea, ONS2004. 42 Steelproducts Offshore A/S Reliability Evaluation of SPO Compact Flange System, Det Norske,

Veritas Report No. 97-3547, Rev. 02, 1997. 43 FABIG Technical Meeting January, 2004. 44 “Static Electricity” – Guidance for plan design engineers” by Graham Hearn, Wolfson

Electrostatics, University of Southampton. 45 NORSOK Standard Z-013, Risk and emergency preparedness analysis, Rev. 2, Annex G;

Procedure for probabilistic explosion simulation. 46 Fire and Blast Information Group, FABIG Technical Note 4, Explosion Resistant Design of

Offshore Structures, TN4, 1996. 47 Response of Secondary Systems to Explosion, S Yasseri, FABIG Newsletter Article No. R454,

FABIG Newsletter Issue No. 33, September 2002. 48 Schlichting H., 1960, Boundary layer Theory, McGraw-Hill, New York. 49 TNO Green Book, Methods for the determination of possible damage, CPR 16E, TNO 1989. 50 Hoerner S. F., 1965, Fluid Dynamics Drag, Published by Author, Midland Park, New Jersey. 51 Baker W E, Cox P A, Westine P S, Kulesz J J and Strehlow R A 1983. Explosion Hazards and

Evaluation, Elsevier Scientific Publishing Company. 52 Kinney and Graham “Explosive Shocks in Air” Elsevier, 1992. 53 Drag Increase due to Vortex Shedding, International Towing Tank Conference, 1981,

Netherlands. 54 The Planning regulations (UK), 1992.



55 Development of a Limit State Approach for Design Against Gas Explosions, Tam V and Corr B,

ERA Major Hazards Offshore Conference, 1999. 56 Shell Private communications, 2002. 57 Gas Explosion Model Evaluation Protocol, 58 Fire and Blast Information Group, FABIG Technical Note 5, Design Guide for Stainless Steel

Blast Walls, 1999. 59 Rotation Capacity of Steel Members Subject to Local Buckling, B Kato: Proc. 9th World

Conference on Earthquake Engineering, Japan 1989 Vol 4. 60 NORSOK Standard Design of Steel Structures, N-004, Revision 1, December 1998. 61 American Society of Civil Engineers (ASCE), ‘Seismic Analysis of Safety Related Nuclear

Structures and Commentary’, 1998. 62 Simplified Methods for Response Analysis to Explosion Loading, FABIG Technical Note 7, The

Steel Construction Institute, 2002. 63 Uncertainties in QRA: How to minimise impact on risk management, E Skramstad, Notes of a

presentation at the FABIG Technical Meeting on Risk Based Design, Fire and Blast Information Group, The Steel Construction Institute, 1999.

64 Fire Load Determination for Offshore Structures, Cook J and Phelps P J, Session 3, paper 2,

Second ERA Conference on structural Design against Extreme Loads, 1993. 65 NORSOK Standard S-001, Technical Safety, Revision 3, 2000. 66 The SFPE Handbook of Fire Protection Engineering, 2nd Edition, The Society of Fire Protection

Engineers (SFPE) and The National Fire Protection Association (NFPA), published by NFPA (USA), 1995.

67 Petrellus internal report on Fire Survivability of Components of Pressurised Process Systems,

2002. 68 Heat Transfer, Ozisik, M Necati, McGraw-Hill Book Company 1985, ISBN -0-07-047982-8. 69 VESSFIRE, An Analysis System for Depressurisation of Process Segments and Process

Equipment Exposed and Not Exposed to Fire, User's Manual, Petrell as, 2002 70 Fire / Blast Performance of Explosion / Fire Damaged Structural and Containment Steelwork,

FR6, 1991. 71 BS 5950-8:1990, Structural Standard for Steel work in Building, Code of Practice for fire

resistant design, 1990. 72 OTI 92 606, Passive Fire Protection: Performance Requirements and Test methods, Health and

Safety Laboratories, 1992. 73 OTI 92 607, Availability and Properties of Passive and Active Fire Protection Systems, Health

and Safety Laboratories, 1992.



74 Controlling Hydrocarbon Fires in Offshore Structures, G. A. Chamberlain, Shell Global

Solutions, Proceedings of the Offshore Technology Conference, May 2002, OTC Paper No. OTC 14132, Houston, Texas, USA.

75 The Register of Type Approved Products No. 8 Fire Safety 1999 – Det Norske Veritas. 76 List of LR Type Approved Products Part 4 – Fire Protection and Fire Fighting 1998, Lloyds

Register. 77 Jet Fire Test Working Group, 1995, Jet Fire Resistance Test of Passive Fire Protection Materials,

HSE Offshore Technology Information Report OTI 95 634. 78 Mather P and Smart R D 1998, Large scale and medium scale jet fire tests, Offshore Technology

Report OTO 97 079. 79 Roberts T, Developments of the Test to Determine the Resistance to Jet Fires of PFP Materials,

FABIG Newsletter Issue No. 35, April 2003, and the FABIG Technical Meeting on Passive Fire Protection Material (July, 2003).

80 BSI Committee FSH/22/14, Determination of the resistance to jet fires of passive fire protection

materials. Formally submitted to the International Standards Committee (ISO committee TC92/SC2/WG8) for conversion to an ISO standard, 2003.

81 Risk-Based Design of Passive Fire Protection, S Yasseri, FABIG Newsletter Article No. R417,

FABIG Newsletter Issue No. 30, October 2001. 82 Roberts, T, HSE Continuation of the Shell Programme on Weathered Passive Fire Protection,

Health and Safety Laboratories, FABIG Technical Meeting on Passive Fire Protection, 1 & 2 July, 2003.

83 Renwick P, Verifying the Suitability of Control and Mitigation Measures, Notes of a presentation

at the FABIG Technical Meeting on Mitigation, Fire and Blast Information Group, The Steel Construction Institute, 2000.

84 Shirvill L and Lowesmith B, Developments in Fire Engineering, FABIG Article NO. R396,

FABIG Newsletter, Issue No. 28, April 2001. 85 NFPA 15, Water Spray Fixed Systems, National Fire Protection Association Standard (am). 86 Medonos S and Geddes P, The Importance of Verifiable Fire Protection and Design, Paper

No. IBP479_04, Rio Oil & Gas 2004, Rio de Janeiro 2004. 87 API Specification 5L, 42nd Edition, January 2000, Specification for Line Pipe. 88 BS EN 10208 Steel Pipes for pipelines for combustible fluids. Technical delivery conditions.

Pipes for requirement Class A. 89 BS EN 10208 Steel Pipes for pipelines for combustible fluids. Technical delivery conditions.

Pipes for requirement Class B. 90 BS EN 10208 Steel Pipes for pipelines for combustible fluids. Technical delivery conditions.

Pipes for requirement Class C. 91 ASTM A106 Standard Specification for seamless carbon steel pipe for high temperature service.



92 ASTM A333 Standard specification for seamless and welded steel pipe for low temperature

service. 93 BPVC Section III Rules for the Construction of Nuclear Power Plant Components – Division 1 –

Subsection NC Class 2 Components and Subsection NC – Class 3 Components. 94 ASCE Structural Analysis and Design of Nuclear Plant Facilities, Committee on Nuclear

Structures ad Materials of the Structure Division, 1980. 95 BS EN 10216-1:2002 Seamless Steel tubes with specified room temperature properties for

pressure purposes – Technical delivery conditions – Part 1: Non-alloy steel tubes with specified room temperature properties.

96 BS EN 10217-1:2002 Welded Steel tubes with specified room temperature properties for pressure

purposes – Technical delivery conditions – Part 1: Non-alloy steel tubes with specified room temperature properties.

97 BS EN 10216-2:2002 Seamless Steel tubes with specified room temperature properties for

pressure purposes – Technical delivery conditions – Part 2: Non-alloy and alloy steel tubes with specified elevated temperature properties.



99 BS EN 10216-4:2002 Seamless Steel tubes with specified room temperature properties for

pressure purposes – Technical delivery conditions – Part 4: Electric welded non-alloy steel tubes with specified low temperature properties.



101 BS 3604-2:1991 Steel pipes and tubes for pressure purposes. Ferritic alloy steel with specified

elevated temperature properties. Specification for longitudinally arc welded tubes. 102 BS 3605-1:1991 Austenitic stainless steel pipes and tubes for pressure purposes. Specification for

seamless tubes. 103 FABIG Technical Meeting - presentation by Jens Holen (Statoil) on the plan for the development

of procedure for probabilistic determination of fire loading, July 2004.

APPENDIX 1 EXAMPLE Sample Calculation. Introduction. Liquid outlet pipes from vessels which contain large inventories of hydrocarbon liquid are designed to withstand blast loads. The piping in this example is routed such that vessel and piping are supported on common steelwork thus eliminating significant differential deflection of steelwork during blast. The blast calculation extends as far as the first pipe support downstream of the ESD valve. To minimise loads on the valve, it is supported with a hold down, line stop and guide on one side plus a hold down and guide on the other. Loads applied to vessel nozzles are a primary concern, allowable nozzle loads need to be incorporated into the vessel design. In the following example calculation the pipe support arrangement minimises the loads applied to the vessel nozzle.

Calculation. The following calculation is based on a typical arrangement. Results are summarised on the stress isometric. This sample calculation is for example only, further expansion and functional cases may be required. Explanatory Notes Blast Wind. 20 kN/m2 blast drag pressure has been applied to the piping in two horizontal directions. No dynamic or shape factors have been applied to the blast pressure. No vertical blast has been applied because the deck is plated. Material Data.

150

Piping Material A333 Gr 6. Tensile 413.7 MPa. Yield 241 MPa Flange Material A350 LF2. Tensile 482 MPa. Yield 248 MPa Bolts Material A193 B7. Tensile 861 MPa. Yield 723 MPa ASME B31.3 Allowable Stress Allowable Sustained Stress Sh = 138.7 MPa (W+P) Allowable Thermal Stress Range = 207 MPa (T) Occasional Allowable Stress = 1.333 Sh = 183 MPa (W+P+Blast) For blast it is common to limit (W+P+Blast) stress to the Yield Stress of 241 MPa. ASME B31.3 separates thermal and sustained stress. This is due to the differences between load controlled conditions, such as weight and pressure, and deformation-controlled conditions, such as thermal expansion and or end displacements. Engineers will be familiar with what are termed primary and secondary stresses. Background and information with regard to this can be found in reference [3]

151

Cases The cases shown are typical when considering blast, they are not exhaustive, further cases such as seismic or two phase flow may be required. CASE 1 (OPE) W+T1+P1 Weight+Thermal+Pressure CASE 2 (SUS) W+P1 Weight+Pressure CASE 3 (SUS) WIN1 X direction Blast CASE 4 (SUS) WIN3 Z direction Blast CASE 5 (OPE) L5=L1+L3 Weight+Thermal+Pressure +X direction Blast CASE 6 (OPE) L6=L1+L4 Weight+Thermal+Pressure +Z direction Blast CASE 7 (OCC) L7=L2+L3 Weight+Pressure+X direction Blast CASE 8 (OCC) L8=L2+L4 Weight+Pressure+Z direction Blast CASE 9 (OPE) L9=L1-L3 Weight+Thermal+Pressure-X direction Blast CASE 10 (OPE) L10=L1-L4 Weight+Thermal+Pressure -Z direction Blast CASE 11 (OCC) L11=L2-L3 Weight+Pressure-X direction Blast CASE 12 (OCC) L12=L2-L4 Weight+Pressure-Z direction Blast CASE 13 (EXP) L13=L1-L2 Thermal Flanges ASME B31.3 [1] does not address flange loadings. A common method for assessing flanges subject to external bending and torsion is given ASME III NC3658. [2]. This is considered appropriate because it is based on the bolt stress where the allowable bolt stress is greater than 20000 psi. Mfs < 3125(Sy/36000)CAb [in lb] Mfd < 6250(Sy/36000)CAb [in lb] Where C= bolt circle diameter [in] Ab = total cross section area of bolts at root of thread. Sq. in Sy = Flange yield strength psi where the value Sy/36000 shell not be greater than unity. Mfs: bending or torsion moment due to weight, thermal expansion, and other sustained loads. Mfd: bending or torsion moment as defined for Mfs but including dynamic loading. In SI units for this example Mfs = 23056 [Nm] Mfd = 46112 [Nm]

152

Discussion of results. Results are reported in the Caesar II output, maximum stress, nozzle and support loads are summarised on the isometric. Maximum flange bending moment is 5904 Nm at node 130.

References. [1] ASME B31.3 Process Piping [2] ASME Section III, DIVISON 1 [3] Companion Guide to ASME Boiler and Pressure Vessel Code Vol 1 Chapter 17. ASME Press. Editor K.R.Rao.

153

C A E S A R I I VERS 4.40 Page 1 Job Description: PROJECT: BLAST EXAMPLE CLIENT : SCI SYSTEM : FROM SEPARARTOR VIA ESDV LINE No: 10"-300# CASES: T1 = All Hot at maximum Design Temp. NOTES: BLAST req'd from vessel to shut-down valves.

154

C A E S A R I I VERS 4.40 Page 2 PIPE DATA ----------------------------------------------------------------------------- ----------------------------------------------------------------------------- From 10 To 20 DZ= .500 m. PIPE Dia= 2,100.000 mm. Wall= 25.400 mm. Insul= .000 mm. Cor= 3.0000 mm. GENERAL T1= 100 C T2= -10 C P1= 2.0000 MPa Mat= (177)A333 6 E= 204,566 MPa v = .292 Density= 7,833.4399 kg./cu.m. RIGID Weight= .00 N. RESTRAINTS Node 10 ANC WIND Wind Shape= .000 ALLOWABLE STRESSES B31.3 (1999) Sc= 138 MPa Sh1= 138 MPa Sh2= 138 MPa Sh3= 138 MPa Sh4= 138 MPa Sh5= 138 MPa Sh6= 138 MPa Sh7= 138 MPa Sh8= 138 MPa Sh9= 138 MPa ----------------------------------------------------------------------------- From 30 To 40 DY= -.133 m. PIPE Dia= 273.050 mm. Wall= 14.000 mm. Insul= .000 mm. GENERAL Fluid= 815.0000000 kg./cu.m. RESTRAINTS Node 30 ANC Cnode 20 WIND Wind Shape= 1.000 ----------------------------------------------------------------------------- From 40 To 50 DY= -.117 m. PIPE Dia= 273.050 mm. Wall= 9.271 mm. Insul= .000 mm. RIGID Weight= 490.00 N. ----------------------------------------------------------------------------- From 50 To 60 DY= -.117 m. GENERAL Mat= (177)A333 6 E= 204,566 MPa v = .292 Density= 7,833.4399 kg./cu.m. RIGID Weight= 490.00 N. ALLOWABLE STRESSES B31.3 (1999) Sc= 138 MPa Sh1= 138 MPa Sh2= 138 MPa Sh3= 138 MPa Sh4= 138 MPa Sh5= 138 MPa Sh6= 138 MPa Sh7= 138 MPa Sh8= 138 MPa Sh9= 138 MPa F1= .90 F2= .90 ----------------------------------------------------------------------------- From 60 To 70 DY= -.381 m. SINGLE FLANGED BEND at "TO" end Radius= 381.000 mm. (LONG) Bend Angle= 90.000 Angle/Node @1= 45.00 69 ----------------------------------------------------------------------------- From 70 To 80 DZ= 3.000 m. BEND at "TO" end Radius= 381.000 mm. (LONG) Bend Angle= 90.000 Angle/Node @1= 45.00 79 Angle/Node @2= .00 78 ----------------------------------------------------------------------------- From 80 To 90 DX= 1.000 m. RESTRAINTS Node 90 X Node 90 Y ----------------------------------------------------------------------------- From 90 To 100 DX= 2.000 m. BEND at "TO" end Radius= 381.000 mm. (LONG) Bend Angle= 90.000 Angle/Node @1= 45.00 99 Angle/Node @2= .00 98 ----------------------------------------------------------------------------- From 100 To 110 DZ= -3.500 m. RESTRAINTS Node 110 X Node 110 Y Node 110 Z ----------------------------------------------------------------------------- From 110 To 120 DZ= -.500 m. ----------------------------------------------------------------------------- From 120 To 130 DZ= -.117 m. RIGID Weight= 490.00 N. ----------------------------------------------------------------------------- From 130 To 140 DZ= -.568 m. GENERAL RIGID Weight= 5,592.00 N. WIND Wind Shape= 1.500 ----------------------------------------------------------------------------- From 140 To 150 DZ= -.117 m. GENERAL RIGID Weight= 490.00 N. WIND Wind Shape= 1.000 ----------------------------------------------------------------------------- From 150 To 160 DZ= -2.000 m. RESTRAINTS Node 160 X Node 160 Y ----------------------------------------------------------------------------- From 160 To 170 DZ= -4.000 m. RESTRAINTS Node 170 Y Node 170 X

155

C A E S A R I I VERS 4.40 Page 3 WIND/WAVE 10 20 WIND Wind Shape= .000 30 40 WIND Wind Shape= 1.000 130 140 WIND Wind Shape= 1.500 140 150 WIND Wind Shape= 1.000

156

C A E S A R I I VERS 4.40 Page 4 INPUT UNITS USED... UNITS= B&RUNIT. NOM/SCH INPUT= ON LENGTH inches x 25.400 = mm. FORCE pounds x 4.448 = N. MASS(dynamics) pounds x 0.454 = kg. MOMENTS(INPUT) inch-pounds x 0.113 = N.m. MOMENTS(OUTPUT) inch-pounds x 0.113 = N.m. STRESS lbs./sq.in. x 0.007 = MPa TEMP. SCALE degrees F. x 0.556 = C PRESSURE psig x 0.007 = MPa ELASTIC MODULUS lbs./sq.in. x 0.007 = MPa PIPE DENSITY lbs./cu.in. x 27680.000 = kg./cu.m. INSULATION DENS. lbs./cu.in. x 27680.000 = kg./cu.m. FLUID DENSITY lbs./cu.in. x 27680.000 = kg./cu.m. TRANSL. STIF lbs./in. x 0.175 = N./mm. ROTATIONAL STIF in.lb./deg. x 0.113 = N.m./deg UNIFORM LOAD lb./in. x 175.120 = N./m G LOAD g's x 1.000 = g's WIND LOAD lbs./sq.in. x 6894.757 = N./sq.m. ELEVATION inches x 0.025 = m. COMPOUND LENGTH inches x 0.025 = m. DIAMETER inches x 25.400 = mm. WALL THICKNESS inches x 25.400 = mm. .

157

C A E S A R I I VERS 4.40 Page 5 COORDINATE REPORT /--------------------(mm.)----------------------/ NODE X Y Z 10 .0000 .0000 .0000 30 .0000 .0000 500.0000 40 .0000 -133.0000 500.0000 50 .0000 -250.0000 500.0000 60 .0000 -367.0000 500.0000 70 .0000 -748.0000 500.0000 80 .0000 -748.0000 3500.0000 90 1000.0000 -748.0000 3500.0000 100 3000.0000 -748.0000 3500.0000 110 3000.0000 -748.0000 .0000 120 3000.0000 -748.0000 -500.0000 130 3000.0000 -748.0000 -617.0000 140 3000.0000 -748.0000 -1185.0000 150 3000.0000 -748.0000 -1302.0000 160 3000.0000 -748.0000 -3302.0000 170 3000.0000 -748.0000 -7302.0000

158

CAESAR II Ver.4.40] Page: 6 RESTRAINT REPORT, Loads on Restraints ... ----------------------- RESTRAINT SUMMARY REPORT LOAD CASE DEFINITION KEY CASE 1 (OPE) W+T1+P1 CASE 2 (SUS) W+P1 CASE 3 (SUS) WIN1 CASE 4 (SUS) WIN3 CASE 5 (OPE) L5=L1+L3 CASE 6 (OPE) L6=L1+L4 CASE 7 (OCC) L7=L2+L3 CASE 8 (OCC) L8=L2+L4 CASE 9 (OPE) L9=L1-L3 CASE 10 (OPE) L10=L1-L4 CASE 11 (OCC) L11=L2-L3 CASE 12 (OCC) L12=L2-L4

159

CAESAR II Ver.4.40 Page: 7 RESTRAINT-DISPLACEMENT REPORT, Loads on Restraints RESTRAINT SUMMARY -----------Forces(N. )---------- ----------Moments(N.m. )--------- --------Displacements(mm.)--------- NODE CASE TYPE FX FY FZ MX MY MZ DX DY DZ 10 Rigid ANC 1 OPE -1134. -2261. 455. -159. -2918. -589. 0.0000 0.0000 0.0000 2 SUS 26. -2726. 25. 1344. 82. -286. 0.0000 0.0000 0.0000 3 SUS 11649. -76. -1392. 1078. 8210. 6636. 0.0000 0.0000 0.0000 4 SUS 726. -389. 12241. -6294. 851. 307. 0.0000 0.0000 0.0000 5 OPE 10515. -2338. -938. 919. 5292. 6047. 0.0000 0.0000 0.0000 6 OPE -407. -2651. 12696. -6453. -2067. -282. 0.0000 0.0000 0.0000 7 OCC 11676. -2802. -1367. 2422. 8292. 6349. 0.0000 0.0000 0.0000 8 OCC 753. -3115. 12266. -4950. 933. 21. 0.0000 0.0000 0.0000 9 OPE -12783. -2185. 1847. -1237. -11127. -7224. 0.0000 0.0000 0.0000 10 OPE -1860. -1872. -11786. 6135. -3768. -896. 0.0000 0.0000 0.0000 11 OCC -11623. -2650. 1418. 266. -8128. -6922. 0.0000 0.0000 0.0000 12 OCC -700. -2337. -12216. 7638. -768. -593. 0.0000 0.0000 0.0000 MAX. 12783./ 9 3115./ 8 12696./ 6 7638./12 11127./ 9 7224./ 9 0.000/ 9 0.000/10 0.000/ 6 30 Rigid ANC 1 OPE -1134. -2261. 455. -1289. -2351. -589. 0.0000 0.0000 0.6091 2 SUS 26. -2726. 25. -19. 69. -286. 0.0000 0.0000 0.0036 3 SUS 11649. -76. -1392. 1040. 2385. 6636. 0.0001 0.0000 0.0000 4 SUS 726. -389. 12241. -6489. 487. 307. 0.0000 0.0000 0.0000 5 OPE 10515. -2338. -938. -250. 34. 6047. 0.0001 0.0000 0.6091 6 OPE -407. -2651. 12696. -7778. -1863. -282. 0.0000 0.0000 0.6092 7 OCC 11676. -2802. -1367. 1020. 2454. 6349. 0.0001 0.0000 0.0036 8 OCC 753. -3115. 12266. -6508. 556. 21. 0.0000 0.0000 0.0037 9 OPE -12783. -2185. 1847. -2329. -4736. -7224. -0.0001 0.0000 0.6091 10 OPE -1860. -1872. -11786. 5199. -2838. -896. 0.0000 0.0000 0.6091 11 OCC -11623. -2650. 1418. -1059. -2316. -6922. -0.0001 0.0000 0.0037 12 OCC -700. -2337. -12216. 6470. -418. -593. 0.0000 0.0000 0.0036 MAX. 12783./ 9 3115./ 8 12696./ 6 7778./ 6 4736./ 9 7224./ 9 0.000/ 9 0.000/ 7 0.609/ 6 90 Rigid X Rigid Y 1 OPE -1341. -5150. 0. 0. 0. 0. 0.0000 0.0000 4.5402 2 SUS -85. -4735. 0. 0. 0. 0. 0.0000 0.0000 0.0915 3 SUS 16636. -120. 0. 0. 0. 0. 0.0000 0.0000 0.0167 4 SUS -19. 466. 0. 0. 0. 0. 0.0000 0.0000 0.3216 5 OPE 15296. -5270. 0. 0. 0. 0. 0.0000 0.0000 4.5568 6 OPE -1360. -4684. 0. 0. 0. 0. 0.0000 0.0000 4.8617 7 OCC 16551. -4855. 0. 0. 0. 0. 0.0000 0.0000 0.1081 8 OCC -104. -4270. 0. 0. 0. 0. 0.0000 0.0000 0.4130 9 OPE -17977. -5029. 0. 0. 0. 0. 0.0000 0.0000 4.5235 10 OPE -1322. -5615. 0. 0. 0. 0. 0.0000 0.0000 4.2186 11 OCC -16722. -4615. 0. 0. 0. 0. 0.0000 0.0000 0.0748 12 OCC -66. -5201. 0. 0. 0. 0. 0.0000 0.0000 -0.2301 MAX. 17977./ 9 5615./10 0./ 1 0./ 1 0./ 1 0./ 1 0.000/ 9 0.000/ 3 4.862/ 6 110 Rigid X Rigid Y Rigid Z 1 OPE 4675. -9898. -455. 0. 0. 0. 0.0000 0.0000 0.0000 2 SUS 109. -9943. -25. 0. 0. 0. 0.0000 0.0000 0.0000 3 SUS 19973. 358. 1392. 0. 0. 0. 0.0000 0.0000 0.0000 4 SUS -947. -173. 8226. 0. 0. 0. 0.0000 0.0000 0.0000 5 OPE 24649. -9540. 938. 0. 0. 0. 0.0000 0.0000 0.0000 6 OPE 3728. -10071. 7771. 0. 0. 0. 0.0000 0.0000 0.0000 7 OCC 20082. -9585. 1367. 0. 0. 0. 0.0000 0.0000 0.0000 8 OCC -839. -10115. 8201. 0. 0. 0. 0.0000 0.0000 0.0000 9 OPE -15298. -10256. -1847. 0. 0. 0. 0.0000 0.0000 0.0000 10 OPE 5622. -9725. -8680. 0. 0. 0. 0.0000 0.0000 0.0000 11 OCC -19865. -10300. -1418. 0. 0. 0. 0.0000 0.0000 0.0000 12 OCC 1056. -9770. -8251. 0. 0. 0. 0.0000 0.0000 0.0000 MAX. 24649./ 5 10300./11 8680./10 0./ 1 0./ 1 0./ 1 0.000/ 5 0.000/11 0.000/10 160 Rigid X Rigid Y 1 OPE -2455. -5139. 0. 0. 0. 0. 0.0000 0.0000 -4.0676 2 SUS -55. -5033. 0. 0. 0. 0. 0.0000 0.0000 -0.0690 3 SUS 23169. -180. 0. 0. 0. 0. 0.0000 0.0000 0.0000 4 SUS 267. 107. 0. 0. 0. 0. 0.0000 0.0000 0.0000 5 OPE 20713. -5319. 0. 0. 0. 0. 0.0000 0.0000 -4.0676 6 OPE -2188. -5032. 0. 0. 0. 0. 0.0000 0.0000 -4.0676 7 OCC 23113. -5213. 0. 0. 0. 0. 0.0000 0.0000 -0.0690 8 OCC 212. -4926. 0. 0. 0. 0. 0.0000 0.0000 -0.0690 9 OPE -25624. -4959. 0. 0. 0. 0. 0.0000 0.0000 -4.0676 10 OPE -2723. -5247. 0. 0. 0. 0. 0.0000 0.0000 -4.0676 11 OCC -23224. -4854. 0. 0. 0. 0. 0.0000 0.0000 -0.0690 12 OCC -323. -5141. 0. 0. 0. 0. 0.0000 0.0000 -0.0690 MAX. 25624./ 9 5319./ 5 0./ 1 0./ 1 0./ 1 0./ 1 0.000/ 9 0.000/ 4 4.068/ 1 170 Rigid Y Rigid X 1 OPE 255. -1530. 0. 0. 0. 0. 0.0000 0.0000 -9.0193 2 SUS 6. -1541. 0. 0. 0. 0. 0.0000 0.0000 -0.1767 3 SUS 8862. 19. 0. 0. 0. 0. 0.0000 0.0000 0.0000 4 SUS -28. -11. 0. 0. 0. 0. 0.0000 0.0000 0.0000 5 OPE 9117. -1512. 0. 0. 0. 0. 0.0000 0.0000 -9.0193 6 OPE 227. -1541. 0. 0. 0. 0. 0.0000 0.0000 -9.0193 7 OCC 8868. -1523. 0. 0. 0. 0. 0.0000 0.0000 -0.1767 8 OCC -22. -1552. 0. 0. 0. 0. 0.0000 0.0000 -0.1767 9 OPE -8608. -1549. 0. 0. 0. 0. 0.0000 0.0000 -9.0193 10 OPE 283. -1519. 0. 0. 0. 0. 0.0000 0.0000 -9.0193 11 OCC -8857. -1560. 0. 0. 0. 0. 0.0000 0.0000 -0.1767

160

12 OCC 34. -1530. 0. 0. 0. 0. 0.0000 0.0000 -0.1767 MAX. 9117./ 5 1560./11 0./ 1 0./ 1 0./ 1 0./ 1 0.000/ 5 0.000/11 9.019/ 1

161

CAESAR II Ver.4.40 Page: 8 STRESS SUMMARY CASE 2 (SUS) W+P1 **** CODE STRESS CHECK PASSED PIPING CODE: B31.3 -1999, August 31, 2001 HIGHEST STRESSES: (MPa ) CODE STRESS %: 23.6 @NODE 110 STRESS: 32.5 ALLOWABLE: 138. BENDING STRESS: 12.2 @NODE 110 TORSIONAL STRESS: 1.0 @NODE 80 AXIAL STRESS: 20.6 @NODE 60 HOOP STRESS: 41.5 @NODE 69 3D MAX INTENSITY: 44.6 @NODE 90

162

CAESAR II Ver.4.40 Page: 9 STRESS REPORT, Stresses on Elements CASE 2 (SUS) W+P1 -------------Stresses(MPa )------------- STRESS ----Stress(MPa )---- ELEMENT AXIAL BENDING TORSION HOOP MAX 3D STRESS INTENSIFICATION CODE ALLOWABLE % NODES STRESS STRESS STRESS STRESS INTENSITY IN-PLANE OUT-PLANE STRESS STRESS 10 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 20 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 30 11.23 0.50 0.06 22.82 25.86 1.000 1.000 11.74 137.90 9. 40 11.21 0.51 -0.06 22.82 25.86 1.000 1.000 11.72 137.90 9. 40 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 50 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 50 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 60 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 60 20.56 1.37 0.10 41.54 44.57 1.904 1.586 21.59 137.90 16. 69 20.45 1.13 -0.38 41.54 44.57 1.904 1.586 21.29 137.90 15. 69 20.45 1.13 0.38 41.54 44.57 1.904 1.586 21.29 137.90 15. 70 20.29 2.36 -0.45 41.54 44.57 1.904 1.586 22.05 137.90 16. 70 20.29 1.24 0.45 41.54 44.57 1.000 1.000 21.53 137.90 16. 78 20.29 0.26 -0.45 41.54 44.57 1.000 1.000 20.54 137.90 15. 78 20.29 0.53 0.45 41.54 44.57 2.483 2.069 20.69 137.90 15. 79 20.29 4.12 0.01 41.54 44.57 2.483 2.069 23.38 137.90 17. 79 20.29 4.12 -0.01 41.54 44.57 2.483 2.069 23.38 137.90 17. 80 20.29 5.81 1.00 41.54 44.62 2.483 2.069 24.65 137.90 18. 80 20.29 2.81 -1.00 41.54 44.61 1.000 1.000 23.10 137.90 17. 90 20.29 6.87 1.00 41.54 44.62 1.000 1.000 27.16 137.90 20. 90 20.27 6.87 -1.00 41.54 44.62 1.000 1.000 27.14 137.90 20. 98 20.27 0.42 1.00 41.54 44.61 1.000 1.000 20.69 137.90 15. 98 20.27 0.90 -1.00 41.54 44.61 2.483 2.069 20.94 137.90 15. 99 20.27 3.04 0.77 41.54 44.59 2.483 2.069 22.55 137.90 16. 99 20.27 3.04 -0.77 41.54 44.59 2.483 2.069 22.55 137.90 16. 100 20.28 4.56 0.00 41.54 44.57 2.483 2.069 23.70 137.90 17. 100 20.28 2.20 0.00 41.54 44.57 1.000 1.000 22.48 137.90 16. 110 20.28 12.25 0.00 41.54 44.57 1.000 1.000 32.52 137.90 24. 110 20.28 12.25 0.00 41.54 44.57 1.000 1.000 32.53 137.90 24. 120 20.28 2.70 0.00 41.54 44.57 1.000 1.000 22.98 137.90 17. 120 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 130 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 130 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 140 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 140 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 150 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 150 20.28 4.01 0.00 41.54 44.57 1.000 1.000 24.29 137.90 18. 160 20.28 5.28 0.00 41.54 44.57 1.000 1.000 25.56 137.90 19. 160 20.28 5.28 0.00 41.54 44.57 1.000 1.000 25.56 137.90 19. 170 20.28 0.00 0.00 41.54 44.57 1.000 1.000 20.28 137.90 15.

163

CAESAR II Ver.4.40 Page: 10 STRESS SUMMARY CASE 7 (OCC) L7=L2+L3 **** CODE STRESS CHECK PASSED PIPING CODE: B31.3 -1999, August 31, 2001 HIGHEST STRESSES: (MPa ) CODE STRESS %: 24.5 @NODE 160 STRESS: 44.9 ALLOWABLE: 183. BENDING STRESS: 24.6 @NODE 160 TORSIONAL STRESS: 3.6 @NODE 60 AXIAL STRESS: 21.9 @NODE 98 HOOP STRESS: 41.5 @NODE 69 3D MAX INTENSITY: 45.7 @NODE 160

164

CAESAR II Ver.4.40 Page: 11 STRESS REPORT, Stresses on Elements CASE 7 (OCC) L7=L2+L3 -------------Stresses(MPa )------------- STRESS ----Stress(MPa )---- ELEMENT AXIAL BENDING TORSION HOOP MAX 3D STRESS INTENSIFICATION CODE ALLOWABLE % NODES STRESS STRESS STRESS STRESS INTENSITY IN-PLANE OUT-PLANE STRESS STRESS 10 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 20 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 30 11.24 11.28 2.15 22.82 26.87 1.000 1.000 22.52 183.41 12. 40 11.22 8.62 -2.15 22.82 26.58 1.000 1.000 19.84 183.41 11. 40 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 50 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 50 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 60 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 60 20.58 11.62 3.58 41.54 45.50 1.904 1.586 29.30 183.41 16. 69 20.27 5.07 -1.50 41.54 44.68 1.904 1.586 24.07 183.41 13. 69 20.27 5.07 1.50 41.54 44.68 1.904 1.586 24.07 183.41 13. 70 20.02 3.11 -1.13 41.54 44.62 1.904 1.586 22.35 183.41 12. 70 20.02 1.75 1.13 41.54 44.62 1.000 1.000 21.77 183.41 12. 78 20.02 1.92 -1.13 41.54 44.62 1.000 1.000 21.95 183.41 12. 78 20.02 4.76 1.13 41.54 44.63 2.483 2.069 23.59 183.41 13. 79 19.09 8.89 -0.70 41.54 44.59 2.483 2.069 25.76 183.41 14. 79 19.09 8.89 0.70 41.54 44.60 2.483 2.069 25.76 183.41 14. 80 18.75 13.72 0.66 41.54 44.60 2.483 2.069 29.03 183.41 16. 80 18.75 5.97 -0.66 41.54 44.59 1.000 1.000 24.72 183.41 13. 90 18.75 8.23 0.66 41.54 44.59 1.000 1.000 26.98 183.41 15. 90 21.89 8.23 -0.66 41.54 44.59 1.000 1.000 30.13 183.41 16. 98 21.89 4.63 0.66 41.54 44.59 1.000 1.000 26.52 183.41 14. 98 21.89 11.45 -0.66 41.54 44.60 2.483 2.069 30.49 183.41 17. 99 21.52 7.79 0.57 41.54 44.59 2.483 2.069 27.37 183.41 15. 99 21.52 7.79 -0.57 41.54 44.58 2.483 2.069 27.37 183.41 15. 100 20.54 6.33 0.00 41.54 44.57 2.483 2.069 25.29 183.41 14. 100 20.54 2.73 0.00 41.54 44.57 1.000 1.000 23.27 183.41 13. 110 20.54 20.39 0.00 41.54 44.57 1.000 1.000 40.93 183.41 22. 110 20.28 20.39 0.00 41.54 44.57 1.000 1.000 40.67 183.41 22. 120 20.28 5.66 0.00 41.54 44.57 1.000 1.000 25.94 183.41 14. 120 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 130 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 130 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 140 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 140 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 150 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 150 20.28 5.79 0.00 41.54 44.57 1.000 1.000 26.07 183.41 14. 160 20.28 24.59 0.00 41.54 45.74 1.000 1.000 44.87 183.41 24. 160 20.28 24.59 0.00 41.54 45.74 1.000 1.000 44.87 183.41 24. 170 20.28 0.00 0.00 41.54 44.57 1.000 1.000 20.28 183.41 11.

165

CAESAR II Ver.4.40 Page: 12 STRESS SUMMARY CASE 8 (OCC) L8=L2+L4 **** CODE STRESS CHECK PASSED PIPING CODE: B31.3 -1999, August 31, 2001 HIGHEST STRESSES: (MPa ) CODE STRESS %: 19.1 @NODE 110 STRESS: 35.0 ALLOWABLE: 183. BENDING STRESS: 13.2 @NODE 60 TORSIONAL STRESS: 1.0 @NODE 80 AXIAL STRESS: 21.8 @NODE 100 HOOP STRESS: 41.5 @NODE 69 3D MAX INTENSITY: 44.7 @NODE 90

166

CAESAR II Ver.4.40 [ Page: 13 STRESS REPORT, Stresses on Elements CASE 8 (OCC) L8=L2+L4 -------------Stresses(MPa )------------- STRESS ----Stress(MPa )---- ELEMENT AXIAL BENDING TORSION HOOP MAX 3D STRESS INTENSIFICATION CODE ALLOWABLE % NODES STRESS STRESS STRESS STRESS INTENSITY IN-PLANE OUT-PLANE STRESS STRESS 10 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 20 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 30 11.28 11.41 0.49 22.82 25.94 1.000 1.000 22.69 183.41 12. 40 11.26 8.64 -0.49 22.82 25.91 1.000 1.000 19.89 183.41 11. 40 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 50 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 50 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 60 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 60 20.64 13.24 0.81 41.54 44.62 1.904 1.586 30.57 183.41 17. 69 21.68 0.10 -0.96 41.54 44.61 1.904 1.586 21.75 183.41 12. 69 21.68 0.10 0.96 41.54 44.60 1.904 1.586 21.75 183.41 12. 70 21.84 3.52 -0.79 41.54 44.60 1.904 1.586 24.48 183.41 13. 70 21.84 1.90 0.79 41.54 44.59 1.000 1.000 23.74 183.41 13. 78 21.84 4.18 -0.79 41.54 44.60 1.000 1.000 26.02 183.41 14. 78 21.84 10.35 0.79 41.54 44.61 2.483 2.069 29.60 183.41 16. 79 21.40 7.46 -0.15 41.54 44.57 2.483 2.069 27.00 183.41 15. 79 21.40 7.46 0.15 41.54 44.57 2.483 2.069 27.00 183.41 15. 80 20.43 9.62 0.99 41.54 44.63 2.483 2.069 27.64 183.41 15. 80 20.43 4.23 -0.99 41.54 44.61 1.000 1.000 24.65 183.41 13. 90 20.43 12.63 0.99 41.54 44.65 1.000 1.000 33.06 183.41 18. 90 20.41 12.63 -0.99 41.54 44.65 1.000 1.000 33.04 183.41 18. 98 20.41 2.86 0.99 41.54 44.61 1.000 1.000 23.27 183.41 13. 98 20.41 7.10 -0.99 41.54 44.62 2.483 2.069 25.73 183.41 14. 99 21.39 6.44 0.74 41.54 44.60 2.483 2.069 26.22 183.41 14. 99 21.39 6.44 -0.74 41.54 44.59 2.483 2.069 26.22 183.41 14. 100 21.84 11.59 0.00 41.54 44.57 2.483 2.069 30.54 183.41 17. 100 21.84 4.81 0.00 41.54 44.57 1.000 1.000 26.65 183.41 15. 110 21.84 13.13 0.00 41.54 44.57 1.000 1.000 34.97 183.41 19. 110 20.28 13.13 0.00 41.54 44.57 1.000 1.000 33.41 183.41 18. 120 20.28 3.58 0.00 41.54 44.57 1.000 1.000 23.86 183.41 13. 120 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 130 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 130 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 140 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 140 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 150 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 150 20.28 3.67 0.00 41.54 44.57 1.000 1.000 23.96 183.41 13. 160 20.28 5.16 0.00 41.54 44.57 1.000 1.000 25.44 183.41 14. 160 20.28 5.16 0.00 41.54 44.57 1.000 1.000 25.44 183.41 14. 170 20.28 0.00 0.00 41.54 44.57 1.000 1.000 20.28 183.41 11.

167

CAESAR II Ver.4.40 Page: 14 STRESS SUMMARY CASE 11 (OCC) L11=L2-L3 **** CODE STRESS CHECK PASSED PIPING CODE: B31.3 -1999, August 31, 2001 HIGHEST STRESSES: (MPa ) CODE STRESS %: 24.5 @NODE 160 STRESS: 44.9 ALLOWABLE: 183. BENDING STRESS: 24.6 @NODE 160 TORSIONAL STRESS: 3.4 @NODE 60 AXIAL STRESS: 21.8 @NODE 80 HOOP STRESS: 41.5 @NODE 69 3D MAX INTENSITY: 45.8 @NODE 160

168

CAESAR II Ver.4.40 Page: 15 STRESS REPORT, Stresses on Elements CASE 11 (OCC) L11=L2-L3 -------------Stresses(MPa )------------- STRESS ----Stress(MPa )---- ELEMENT AXIAL BENDING TORSION HOOP MAX 3D STRESS INTENSIFICATION CODE ALLOWABLE % NODES STRESS STRESS STRESS STRESS INTENSITY IN-PLANE OUT-PLANE STRESS STRESS 10 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 20 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 30 11.22 12.28 -2.03 22.82 27.00 1.000 1.000 23.50 183.41 13. 40 11.21 9.63 2.03 22.82 26.63 1.000 1.000 20.84 183.41 11. 40 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 50 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 50 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 60 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 60 20.55 14.32 -3.38 41.54 45.61 1.904 1.586 31.29 183.41 17. 69 20.62 6.81 0.74 41.54 44.60 1.904 1.586 25.73 183.41 14. 69 20.62 6.81 -0.74 41.54 44.60 1.904 1.586 25.73 183.41 14. 70 20.55 3.21 0.24 41.54 44.57 1.904 1.586 22.96 183.41 13. 70 20.55 1.88 -0.24 41.54 44.57 1.000 1.000 22.43 183.41 12. 78 20.55 2.08 0.24 41.54 44.57 1.000 1.000 22.63 183.41 12. 78 20.55 5.02 -0.24 41.54 44.57 2.483 2.069 24.32 183.41 13. 79 21.49 7.95 0.72 41.54 44.60 2.483 2.069 27.45 183.41 15. 79 21.49 7.95 -0.72 41.54 44.59 2.483 2.069 27.45 183.41 15. 80 21.83 11.25 1.33 41.54 44.72 2.483 2.069 30.27 183.41 17. 80 21.83 4.61 -1.33 41.54 44.65 1.000 1.000 26.44 183.41 14. 90 21.83 5.99 1.33 41.54 44.67 1.000 1.000 27.82 183.41 15. 90 18.65 5.99 -1.33 41.54 44.67 1.000 1.000 24.64 183.41 13. 98 18.65 4.91 1.33 41.54 44.65 1.000 1.000 23.56 183.41 13. 98 18.65 12.20 -1.33 41.54 44.74 2.483 2.069 27.80 183.41 15. 99 19.02 9.25 0.96 41.54 44.62 2.483 2.069 25.95 183.41 14. 99 19.02 9.25 -0.96 41.54 44.63 2.483 2.069 25.95 183.41 14. 100 20.01 7.15 0.00 41.54 44.57 2.483 2.069 25.38 183.41 14. 100 20.01 3.23 0.00 41.54 44.57 1.000 1.000 23.24 183.41 13. 110 20.01 21.29 0.00 41.54 44.57 1.000 1.000 41.30 183.41 23. 110 20.28 21.29 0.00 41.54 44.57 1.000 1.000 41.57 183.41 23. 120 20.28 6.07 0.00 41.54 44.57 1.000 1.000 26.36 183.41 14. 120 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 130 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 130 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 140 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 140 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 150 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 150 20.28 5.01 0.00 41.54 44.57 1.000 1.000 25.29 183.41 14. 160 20.28 24.63 0.00 41.54 45.78 1.000 1.000 44.91 183.41 24. 160 20.28 24.63 0.00 41.54 45.78 1.000 1.000 44.91 183.41 24. 170 20.28 0.00 0.00 41.54 44.57 1.000 1.000 20.28 183.41 11.

169

CAESAR II Ver.4.40 Page: 16 STRESS SUMMARY CASE 12 (OCC) L12=L2-L4 **** CODE STRESS CHECK PASSED PIPING CODE: B31.3 -1999, August 31, 2001 HIGHEST STRESSES: (MPa ) CODE STRESS %: 18.1 @NODE 90 STRESS: 33.2 ALLOWABLE: 183. BENDING STRESS: 13.2 @NODE 60 TORSIONAL STRESS: 1.0 @NODE 80 AXIAL STRESS: 20.5 @NODE 60 HOOP STRESS: 41.5 @NODE 69 3D MAX INTENSITY: 44.7 @NODE 90

170

CAESAR II Ver.4.40 Page: 17 STRESS REPORT, Stresses on Elements CASE 12 (OCC) L12=L2-L4 -------------Stresses(MPa )------------- STRESS ----Stress(MPa )---- ELEMENT AXIAL BENDING TORSION HOOP MAX 3D STRESS INTENSIFICATION CODE ALLOWABLE % NODES STRESS STRESS STRESS STRESS INTENSITY IN-PLANE OUT-PLANE STRESS STRESS 10 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 20 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 30 11.19 11.39 -0.37 22.82 25.91 1.000 1.000 22.58 183.41 12. 40 11.17 8.62 0.37 22.82 25.89 1.000 1.000 19.80 183.41 11. 40 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 50 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 50 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 60 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 60 20.49 13.17 -0.61 41.54 44.60 1.904 1.586 30.37 183.41 17. 69 19.21 2.35 0.20 41.54 44.57 1.904 1.586 20.98 183.41 11. 69 19.21 2.35 -0.20 41.54 44.57 1.904 1.586 20.98 183.41 11. 70 18.74 8.00 -0.10 41.54 44.57 1.904 1.586 24.74 183.41 13. 70 18.74 4.21 0.10 41.54 44.57 1.000 1.000 22.95 183.41 13. 78 18.74 4.13 -0.10 41.54 44.57 1.000 1.000 22.87 183.41 12. 78 18.74 10.26 0.10 41.54 44.57 2.483 2.069 26.43 183.41 14. 79 19.18 6.22 0.17 41.54 44.57 2.483 2.069 23.84 183.41 13. 79 19.18 6.22 -0.17 41.54 44.57 2.483 2.069 23.84 183.41 13. 80 20.15 8.95 1.00 41.54 44.63 2.483 2.069 26.87 183.41 15. 80 20.15 3.87 -1.00 41.54 44.61 1.000 1.000 24.02 183.41 13. 90 20.15 13.04 1.00 41.54 44.66 1.000 1.000 33.19 183.41 18. 90 20.14 13.04 -1.00 41.54 44.66 1.000 1.000 33.17 183.41 18. 98 20.14 2.57 1.00 41.54 44.61 1.000 1.000 22.70 183.41 12. 98 20.14 6.34 -1.00 41.54 44.62 2.483 2.069 24.89 183.41 14. 99 19.15 7.19 0.79 41.54 44.60 2.483 2.069 24.55 183.41 13. 99 19.15 7.19 -0.79 41.54 44.60 2.483 2.069 24.55 183.41 13. 100 18.71 12.33 0.00 41.54 44.57 2.483 2.069 27.96 183.41 15. 100 18.71 5.12 0.00 41.54 44.57 1.000 1.000 23.84 183.41 13. 110 18.71 11.69 0.00 41.54 44.57 1.000 1.000 30.40 183.41 17. 110 20.28 11.69 0.00 41.54 44.57 1.000 1.000 31.97 183.41 17. 120 20.28 2.82 0.00 41.54 44.57 1.000 1.000 23.11 183.41 13. 120 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 130 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 130 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 140 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 140 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 150 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 150 20.28 4.62 0.00 41.54 44.57 1.000 1.000 24.91 183.41 14. 160 20.28 5.42 0.00 41.54 44.57 1.000 1.000 25.71 183.41 14. 160 20.28 5.42 0.00 41.54 44.57 1.000 1.000 25.71 183.41 14. 170 20.28 0.00 0.00 41.54 44.57 1.000 1.000 20.28 183.41 11.

171

CAESAR II Ver.4.40 Page: 18 STRESS SUMMARY CASE 13 (EXP) L13=L1-L2 **** CODE STRESS CHECK PASSED PIPING CODE: B31.3 -1999, August 31, 2001 HIGHEST STRESSES: (MPa ) CODE STRESS %: 6.7 @NODE 110 STRESS: 12.5 ALLOWABLE: 186. BENDING STRESS: 12.5 @NODE 110 TORSIONAL STRESS: 2.5 @NODE 60 AXIAL STRESS: 0.3 @NODE 98 HOOP STRESS: 0.0 @NODE 20 3D MAX INTENSITY: 17.9 @NODE 110

172

CAESAR II Ver.4.40 Page: 19 STRESS REPORT, Stresses on Elements CASE 13 (EXP) L13=L1-L2 -------------Stresses(MPa )------------- STRESS ----Stress(MPa )---- ELEMENT AXIAL BENDING TORSION HOOP MAX 3D STRESS INTENSIFICATION CODE ALLOWABLE % NODES STRESS STRESS STRESS STRESS INTENSITY IN-PLANE OUT-PLANE STRESS STRESS 10 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 20 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 30 -0.04 1.86 -1.72 0.00 4.17 1.000 1.000 3.92 206.85 2. 40 -0.04 1.74 1.72 0.00 4.09 1.000 1.000 3.86 206.85 2. 40 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 50 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 50 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 60 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 60 -0.06 4.34 -2.47 0.00 8.00 1.904 1.586 6.57 186.17 4. 69 0.00 5.61 1.97 0.00 8.94 1.904 1.586 6.85 186.17 4. 69 0.00 5.61 -1.97 0.00 8.94 1.904 1.586 6.85 186.17 4. 70 0.06 7.07 0.58 0.00 10.26 1.904 1.586 7.16 186.17 4. 70 0.06 4.33 -0.58 0.00 6.38 1.000 1.000 4.48 186.17 2. 78 0.06 1.38 0.58 0.00 2.35 1.000 1.000 1.80 186.17 1. 78 0.06 3.33 -0.58 0.00 4.99 2.483 2.069 3.53 186.17 2. 79 -0.07 4.98 0.65 0.00 7.33 2.483 2.069 5.15 186.17 3. 79 -0.07 4.98 -0.65 0.00 7.33 2.483 2.069 5.15 186.17 3. 80 -0.15 6.42 0.45 0.00 9.44 2.483 2.069 6.48 186.17 3. 80 -0.15 2.62 -0.45 0.00 4.07 1.000 1.000 2.77 186.17 1. 90 -0.15 3.05 0.45 0.00 4.67 1.000 1.000 3.18 186.17 2. 90 -0.31 3.05 -0.45 0.00 4.90 1.000 1.000 3.18 186.17 2. 98 -0.31 4.46 0.45 0.00 6.90 1.000 1.000 4.55 186.17 2. 98 -0.31 11.08 -0.45 0.00 16.32 2.483 2.069 11.11 186.17 6. 99 -0.26 10.38 0.31 0.00 15.24 2.483 2.069 10.40 186.17 6. 99 -0.26 10.38 -0.31 0.00 15.24 2.483 2.069 10.40 186.17 6. 100 -0.06 7.46 0.00 0.00 10.75 2.483 2.069 7.46 186.17 4. 100 -0.06 3.04 0.00 0.00 4.43 1.000 1.000 3.04 186.17 2. 110 -0.06 12.47 0.00 0.00 17.92 1.000 1.000 12.47 186.17 7. 110 0.00 12.47 0.00 0.00 17.84 1.000 1.000 12.47 186.17 7. 120 0.00 10.28 0.00 0.00 14.69 1.000 1.000 10.28 186.17 6. 120 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 130 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 130 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 140 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 140 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 150 0.00 0.00 0.00 0.00 0.00 0.000 0.000 0.00 0.00 0. 150 0.00 6.75 0.00 0.00 9.65 1.000 1.000 6.75 186.17 4. 160 0.00 2.04 0.00 0.00 2.91 1.000 1.000 2.04 186.17 1. 160 0.00 2.04 0.00 0.00 2.91 1.000 1.000 2.04 186.17 1. 170 0.00 0.00 0.00 0.00 0.00 1.000 1.000 0.00 186.17 0.

173

CAESAR II Ver.4.40 Page: 20 FORCE REPORT, Forces on Elements CASE 1 (OPE) W+T1+P1 ELEMENT AXIAL SHEAR BENDING TORSION -----------Forces(N. )----------- ---------Moments(N.m. )------ NODES: FORCE FORCE MOMENT MOMENT FX FY FZ MX MY MZ 10 -455. 2530. 2922.0 588.9 1134. 2261. -455. 158.8 2917.7 588.9 20 455. 2530. 2681.2 -588.9 -1134. -2261. 455. -1289.4 -2350.7 -588.9 30 -2261. 1222. 1417.6 -2350.7 1134. 2261. -455. 1289.4 2350.7 588.9 40 2095. 1222. 1304.7 2350.7 -1134. -2095. 455. -1228.9 -2350.7 -438.1 40 -2095. 1222. 1304.7 -2350.7 1134. 2095. -455. 1228.9 2350.7 438.1 50 1557. 1222. 1214.8 2350.7 -1134. -1557. 455. -1175.7 -2350.7 -305.5 50 -1557. 1222. 1214.8 -2350.7 1134. 1557. -455. 1175.7 2350.7 305.5 60 1020. 1222. 1135.8 2350.7 -1134. -1020. 455. -1122.6 -2350.7 -172.8 60 -1020. 1222. 1135.8 -2350.7 1134. 1020. -455. 1122.6 2350.7 172.8 69 832. 1149. 1838.3 1666.6 -1134. -721. 455. -1091.9 -2224.2 132.7 69 -832. 1149. 1838.3 -1666.6 1134. 721. -455. 1091.9 2224.2 -132.7 70 455. 1210. 2259.4 259.2 -1134. -423. 455. -1193.1 -1918.7 259.2 70 -455. 1210. 2259.4 -259.2 1134. 423. -455. 1193.1 1918.7 -259.2 78 455. 2134. 714.0 259.2 -1134. 1808. 455. 356.1 618.8 259.2 78 -455. 2134. 714.0 -259.2 1134. -1808. -455. -356.1 -618.8 -259.2 79 -480. 2387. 1143.8 648.1 -1134. 2106. 455. 881.0 975.1 35.5 79 480. 2387. 1143.8 -648.1 1134. -2106. -455. -881.0 -975.1 -35.5 80 -1134. 2447. 1352.1 1127.4 -1134. 2404. 455. 1127.4 1224.1 -574.2 80 1134. 2447. 1352.1 -1127.4 1134. -2404. -455. -1127.4 -1224.1 574.2 90 -1134. 3055. 2710.0 1127.4 -1134. 3021. 455. 1127.4 1505.5 -2253.3 90 2475. 2177. 2710.0 -1127.4 2475. 2128. -455. -1127.4 -1505.5 2253.3 98 -2475. 687. 2244.6 1127.4 -2475. -515. 455. 1127.4 2241.7 -113.5 98 2475. 687. 2244.6 -1127.4 2475. 515. -455. -1127.4 -2241.7 113.5 99 -2071. 1445. 2240.8 831.3 -2475. -217. 455. 1162.9 2088.0 -12.8 99 2071. 1445. 2240.8 -831.3 2475. 217. -455. -1162.9 -2088.0 12.8 100 -455. 2476. 1885.9 0.0 -2475. 82. 455. 1178.8 1472.1 0.0 100 455. 2476. 1885.9 0.0 2475. -82. -455. -1178.8 -1472.1 0.0 110 -455. 4038. 7376.9 0.0 -2475. 3191. 455. -3924.4 -6246.4 0.0 110 0. 7059. 7376.9 0.0 -2201. 6708. 0. 3924.4 6246.4 0.0 120 0. 6588. 5192.8 0.0 2201. -6209. 0. -695.2 -5146.1 0.0 120 0. 6588. 5192.8 0.0 -2201. 6209. 0. 695.2 5146.1 0.0 130 0. 6083. 4888.6 0.0 2201. -5672. 0. -0.2 -4888.6 0.0 130 0. 6083. 4888.6 0.0 -2201. 5672. 0. 0.2 4888.6 0.0 140 0. 2206. 3962.0 0.0 2201. 151. 0. 1567.6 -3638.7 0.0 140 0. 2206. 3962.0 0.0 -2201. -151. 0. -1567.6 3638.7 0.0 150 0. 2306. 3706.6 0.0 2201. 689. 0. 1518.4 -3381.3 0.0 150 0. 2306. 3706.6 0.0 -2201. -689. 0. -1518.4 3381.3 0.0 160 0. 3470. 2115.0 0.0 2201. 2682. 0. -1852.9 1019.8 0.0 160 0. 2470. 2115.0 0.0 255. 2457. 0. 1852.9 -1019.8 0.0 170 0. 1551. 0.0 0.0 -255. 1530. 0. 0.0 0.0 0.0

174

CAESAR II Ver.4.40 Page: 21 FORCE REPORT, Forces on Elements CASE 2 (SUS) W+P1 ELEMENT AXIAL SHEAR BENDING TORSION -----------Forces(N. )----------- ---------Moments(N.m. )------ NODES: FORCE FORCE MOMENT MOMENT FX FY FZ MX MY MZ 10 -25. 2726. 1346.2 286.3 -26. 2726. -25. -1343.7 -82.1 286.3 20 25. 2726. 71.6 -286.3 26. -2726. 25. -19.3 68.9 -286.3 30 -2726. 36. 286.9 68.9 -26. 2726. -25. 19.3 -68.9 286.3 40 2559. 36. 290.2 -68.9 26. -2559. 25. -15.9 68.9 -289.8 40 -2559. 36. 290.2 68.9 -26. 2559. -25. 15.9 -68.9 289.8 50 2022. 36. 293.1 -68.9 26. -2022. 25. -13.0 68.9 -292.9 50 -2022. 36. 293.1 68.9 -26. 2022. -25. 13.0 -68.9 292.9 60 1484. 36. 296.1 -68.9 26. -1484. 25. -10.0 68.9 -295.9 60 -1484. 36. 296.1 68.9 -26. 1484. -25. 10.0 -68.9 295.9 69 856. 821. 222.9 -261.0 26. -1186. 25. -146.9 66.0 -303.0 69 -856. 821. 222.9 261.0 -26. 1186. -25. 146.9 -66.0 303.0 70 25. 888. 425.4 -306.0 26. -888. 25. -421.3 58.9 -306.0 70 -25. 888. 425.4 306.0 -26. 888. -25. 421.3 -58.9 306.0 78 25. 1343. 88.0 -306.0 26. 1343. 25. 88.0 -0.1 -306.0 78 -25. 1343. 88.0 306.0 -26. -1343. -25. -88.0 0.1 306.0 79 36. 1641. 682.8 7.0 26. 1641. 25. 487.8 -4.4 -477.8 79 -36. 1641. 682.8 -7.0 -26. -1641. -25. -487.8 4.4 477.8 80 26. 1940. 962.3 682.2 26. 1939. 25. 682.2 -0.5 -962.3 80 -26. 1940. 962.3 -682.2 -26. -1939. -25. -682.2 0.5 962.3 90 26. 2557. 2353.8 682.2 26. 2556. 25. 682.2 15.0 -2353.8 90 59. 2179. 2353.8 -682.2 59. 2179. -25. -682.2 -15.0 2353.8 98 -59. 566. 143.9 682.2 -59. -565. 25. 682.2 55.8 -132.7 98 59. 566. 143.9 -682.2 59. 565. -25. -682.2 -55.8 132.7 99 -59. 268. 501.6 524.5 -59. -267. 25. 723.3 56.0 -18.4 99 59. 268. 501.6 -524.5 59. 267. -25. -723.3 -56.0 18.4 100 -25. 67. 754.0 0.0 -59. 31. 25. 752.8 42.9 0.0 100 25. 67. 754.0 0.0 59. -31. -25. -752.8 -42.9 0.0 110 -25. 3141. 4196.0 0.0 -59. 3140. 25. -4193.6 -140.9 0.0 110 0. 6803. 4196.0 0.0 -50. 6802. 0. 4193.6 140.9 0.0 120 0. 6304. 924.4 0.0 50. -6304. 0. -917.0 -116.1 0.0 120 0. 6304. 924.4 0.0 -50. 6304. 0. 917.0 116.1 0.0 130 0. 5767. 238.0 0.0 50. -5766. 0. -210.9 -110.3 0.0 130 0. 5767. 238.0 0.0 -50. 5766. 0. 210.9 110.3 0.0 140 0. 75. 1413.1 0.0 50. 57. 0. 1410.7 -82.1 0.0 140 0. 75. 1413.1 0.0 -50. -57. 0. -1410.7 82.1 0.0 150 0. 596. 1374.8 0.0 50. 594. 0. 1372.7 -76.3 0.0 150 0. 596. 1374.8 0.0 -50. -594. 0. -1372.7 76.3 0.0 160 0. 2588. 1809.1 0.0 50. 2588. 0. -1809.0 23.0 0.0 160 0. 2446. 1809.1 0.0 6. 2446. 0. 1809.0 -23.0 0.0 170 0. 1541. 0.0 0.0 -6. 1541. 0. 0.0 0.0 0.0

175

CAESAR II Ver.4.40 Page: 22 FORCE REPORT, Forces on Elements CASE 5 (OPE) L5=L1+L3 ELEMENT AXIAL SHEAR BENDING TORSION -----------Forces(N. )----------- ---------Moments(N.m. )------ NODES: FORCE FORCE MOMENT MOMENT FX FY FZ MX MY MZ 10 938. 10772. 5371.3 -6046.6 -10515. 2338. 938. -919.0 -5292.1 -6046.6 20 -938. 10772. 252.2 6046.6 10515. -2338. -938. -249.8 34.3 6046.6 30 -2338. 10557. 6051.7 34.3 -10515. 2338. 938. 249.8 -34.3 -6046.6 40 2171. 9834. 4711.2 -34.3 9789. -2171. -938. -374.5 34.3 4696.3 40 -2171. 9834. 4711.2 34.3 -9789. 2171. 938. 374.5 -34.3 -4696.3 50 1634. 9198. 3620.9 -34.3 9150. -1634. -938. -484.3 34.3 3588.4 50 -1634. 9198. 3620.9 34.3 -9150. 1634. 938. 484.3 -34.3 -3588.4 60 1096. 8563. 2623.3 -34.3 8511. -1096. -938. -594.0 34.3 2555.2 60 -1096. 8563. 2623.3 34.3 -8511. 1096. 938. 594.0 -34.3 -2555.2 69 -99. 6898. 971.6 898.1 6788. -798. -938. -947.0 -788.6 481.6 69 99. 6898. 971.6 -898.1 -6788. 798. 938. 947.0 788.6 -481.6 70 -938. 5089. 2669.6 -210.4 5064. -499. -938. -1224.2 -2372.4 -210.4 70 938. 5089. 2669.6 210.4 -5064. 499. 938. 1224.2 2372.4 210.4 78 -938. 7363. 157.2 -210.4 -7157. 1731. -938. 154.2 -30.6 -210.4 78 938. 7363. 157.2 210.4 7157. -1731. 938. -154.2 30.6 210.4 79 -6764. 5804. 2123.4 164.7 -8628. 2029. -938. 658.5 1980.2 -425.6 79 6764. 5804. 2123.4 -164.7 8628. -2029. 938. -658.5 -1980.2 425.6 80 -9237. 2510. 2897.1 896.4 -9237. 2328. -938. 896.4 2713.6 -1014.7 80 9237. 2510. 2897.1 -896.4 9237. -2328. 938. -896.4 -2713.6 1014.7 90 -9237. 3090. 3399.1 896.4 -9237. 2945. -938. 896.4 2133.0 -2646.6 90 -6058. 2507. 3399.1 -896.4 -6058. 2325. 938. -896.4 -2133.0 2646.6 98 6058. 1177. 643.0 896.4 6058. -711. -938. 896.4 614.8 -188.5 98 -6058. 1177. 643.0 -896.4 -6058. 711. 938. -896.4 -614.8 188.5 99 4516. 3216. 1187.0 699.0 5449. -413. -938. 953.8 993.3 -34.8 99 -4516. 3216. 1187.0 -699.0 -5449. 413. 938. -953.8 -993.3 34.8 100 938. 3979. 2378.7 0.0 3978. -115. -938. 1022.7 2147.6 0.0 100 -938. 3979. 2378.7 0.0 -3978. 115. 938. -1022.7 -2147.6 0.0 110 938. 13393. 12497.8 0.0 -13054. 2994. -938. -3467.2 -12007.3 0.0 110 0. 13315. 12497.8 0.0 -11594. 6546. 0. 3467.2 12007.3 0.0 120 0. 10731. 6900.1 0.0 8864. -6048. 0. -318.5 -6892.7 0.0 120 0. 10731. 6900.1 0.0 -8864. 6048. 0. 318.5 6892.7 0.0 130 0. 9900. 5903.9 0.0 8225. -5510. 0. 357.7 -5893.0 0.0 130 0. 9900. 5903.9 0.0 -8225. 5510. 0. -357.7 5893.0 0.0 140 0. 3586. 3134.9 0.0 3573. 312. 0. 1833.9 -2542.5 0.0 140 0. 3586. 3134.9 0.0 -3573. -312. 0. -1833.9 2542.5 0.0 150 0. 3054. 2791.4 0.0 2934. 850. 0. 1765.9 -2161.9 0.0 150 0. 3054. 2791.4 0.0 -2934. -850. 0. -1765.9 2161.9 0.0 160 0. 8479. 7469.1 0.0 -7988. 2843. 0. -1927.6 -7216.1 0.0 160 0. 12964. 7469.1 0.0 -12725. 2475. 0. 1927.6 7216.1 0.0 170 0. 9242. 0.0 0.0 -9117. 1512. 0. 0.0 0.0 0.0

176

CAESAR II Ver.4.40 Page: 23 FORCE REPORT, Forces on Elements CASE 6 (OPE) L6=L1+L4 ELEMENT AXIAL SHEAR BENDING TORSION -----------Forces(N. )----------- ---------Moments(N.m. )------ NODES: FORCE FORCE MOMENT MOMENT FX FY FZ MX MY MZ 10 -12696. 2682. 6775.9 282.1 407. 2651. -12696. 6452.9 2067.1 282.1 20 12696. 2682. 7998.3 -282.1 -407. -2651. 12696. -7778.2 -1863.4 -282.1 30 -2651. 12702. 7783.3 -1863.4 407. 2651. -12696. 7778.2 1863.4 282.1 40 2484. 11976. 6142.2 1863.4 -407. -2484. 11969. -6138.0 -1863.4 -227.9 40 -2484. 11976. 6142.2 -1863.4 407. 2484. -11969. 6138.0 1863.4 227.9 50 1946. 11338. 4778.3 1863.4 -407. -1946. 11331. -4774.9 -1863.4 -180.2 50 -1946. 11338. 4778.3 -1863.4 407. 1946. -11331. 4774.9 1863.4 180.2 60 1409. 10699. 3489.1 1863.4 -407. -1409. 10692. -3486.6 -1863.4 -132.5 60 -1409. 10699. 3489.1 -1863.4 407. 1409. -10692. 3486.6 1863.4 132.5 69 7305. 5749. 1599.0 1269.4 -407. -1111. 9220. -928.8 -1817.9 -22.8 69 -7305. 5749. 1599.0 -1269.4 407. 1111. -9220. 928.8 1817.9 22.8 70 8611. 909. 1717.6 22.7 -407. -812. 8611. -179.8 -1708.2 22.7 70 -8611. 909. 1717.6 -22.7 407. 812. -8611. 179.8 1708.2 -22.7 78 8611. 1476. 939.4 22.7 -407. 1418. 8611. 498.4 -796.3 22.7 78 -8611. 1476. 939.4 -22.7 407. -1418. -8611. -498.4 796.3 -22.7 79 5370. 6189. 794.7 538.1 -407. 1717. 8002. 918.5 229.6 -157.5 79 -5370. 6189. 794.7 -538.1 407. -1717. -8002. -918.5 -229.6 157.5 80 -407. 6834. 2318.4 1121.4 -407. 2015. 6531. 1121.4 2221.7 -662.4 80 407. 6834. 2318.4 -1121.4 407. -2015. -6531. -1121.4 -2221.7 662.4 90 -407. 4105. 5624.9 1121.4 -407. 2632. 3150. 1121.4 5218.0 -2100.6 90 1767. 3760. 5624.9 -1121.4 1767. 2052. -3150. -1121.4 -5218.0 2100.6 98 -1767. 5707. 3162.8 1121.4 -1767. -438. -5691. 1121.4 3161.7 -84.5 98 1767. 5707. 3162.8 -1121.4 1767. 438. 5691. -1121.4 -3161.7 84.5 99 3815. 6315. 1484.0 815.0 -1767. -140. -7162. 1148.3 1244.1 -4.3 99 -3815. 6315. 1484.0 -815.0 1767. 140. 7162. -1148.3 -1244.1 4.3 100 7771. 1774. 1145.0 0.0 -1767. 158. -7771. 1143.7 -54.3 0.0 100 -7771. 1774. 1145.0 0.0 1767. -158. 7771. -1143.7 54.3 0.0 110 7771. 3714. 6971.4 0.0 -1767. 3267. -7771. -4197.6 -5566.0 0.0 110 0. 7081. 6971.4 0.0 -1961. 6804. 0. 4197.6 5566.0 0.0 120 0. 6603. 4677.0 0.0 1961. -6305. 0. -920.3 -4585.6 0.0 120 0. 6603. 4677.0 0.0 -1961. 6305. 0. 920.3 4585.6 0.0 130 0. 6092. 4361.4 0.0 1961. -5768. 0. -214.0 -4356.1 0.0 130 0. 6092. 4361.4 0.0 -1961. 5768. 0. 214.0 4356.1 0.0 140 0. 1962. 3535.1 0.0 1961. 55. 0. 1408.4 -3242.4 0.0 140 0. 1962. 3535.1 0.0 -1961. -55. 0. -1408.4 3242.4 0.0 150 0. 2048. 3310.0 0.0 1961. 593. 0. 1370.5 -3013.0 0.0 150 0. 2048. 3310.0 0.0 -1961. -593. 0. -1370.5 3013.0 0.0 160 0. 3245. 2023.8 0.0 1961. 2586. 0. -1808.3 908.7 0.0 160 0. 2456. 2023.8 0.0 227. 2446. 0. 1808.3 -908.7 0.0 170 0. 1558. 0.0 0.0 -227. 1541. 0. 0.0 0.0 0.0

177

CAESAR II Ver.4.40 Page: 24 FORCE REPORT, Forces on Elements CASE 7 (OCC) L7=L2+L3 ELEMENT AXIAL SHEAR BENDING TORSION -----------Forces(N. )----------- ---------Moments(N.m. )------ NODES: FORCE FORCE MOMENT MOMENT FX FY FZ MX MY MZ 10 1367. 12007. 8638.2 -6349.2 -11676. 2802. 1367. -2421.5 -8291.9 -6349.2 20 -1367. 12007. 2657.7 6349.2 11676. -2802. -1367. 1020.3 2454.0 6349.2 30 -2802. 11755. 6430.7 2454.0 -11676. 2802. 1367. -1020.3 -2454.0 -6349.2 40 2636. 11034. 4916.7 -2454.0 10949. -2636. -1367. 838.5 2454.0 4844.7 40 -2636. 11034. 4916.7 2454.0 -10949. 2636. 1367. -838.5 -2454.0 -4844.7 50 2098. 10401. 3664.3 -2454.0 10310. -2098. -1367. 678.5 2454.0 3601.0 50 -2098. 10401. 3664.3 2454.0 -10310. 2098. 1367. -678.5 -2454.0 -3601.0 60 1561. 9768. 2486.7 -2454.0 9671. -1561. -1367. 518.5 2454.0 2432.0 60 -1561. 9768. 2486.7 2454.0 -9671. 1561. 1367. -518.5 -2454.0 -2432.0 69 -74. 8163. 1094.2 -1029.4 7948. -1262. -1367. -2.0 1501.6 45.9 69 74. 8163. 1094.2 1029.4 -7948. 1262. 1367. 2.0 -1501.6 -45.9 70 -1367. 6299. 600.3 -775.6 6224. -964. -1367. -452.3 -394.7 -775.6 70 1367. 6299. 600.3 775.6 -6224. 964. 1367. 452.3 394.7 775.6 78 -1367. 6129. 659.4 -775.6 -5997. 1267. -1367. -113.9 -649.5 -775.6 78 1367. 6129. 659.4 775.6 5997. -1267. 1367. 113.9 649.5 775.6 79 -6247. 4589. 1314.0 -476.3 -7468. 1565. -1367. 265.3 1000.8 -938.9 79 6247. 4589. 1314.0 476.3 7468. -1565. 1367. -265.3 -1000.8 938.9 80 -8077. 2311. 2045.7 451.2 -8077. 1863. -1367. 451.2 1488.9 -1402.9 80 8077. 2311. 2045.7 -451.2 8077. -1863. 1367. -451.2 -1488.9 1402.9 90 -8077. 2832. 2821.2 451.2 -8077. 2480. -1367. 451.2 642.6 -2747.1 90 -8474. 2741. 2821.2 -451.2 -8474. 2375. 1367. -451.2 -642.6 2747.1 98 8474. 1565. 1584.8 451.2 8474. -762. -1367. 451.2 -1571.1 -207.6 98 -8474. 1565. 1584.8 -451.2 -8474. 762. 1367. -451.2 1571.1 207.6 99 6528. 4618. 1091.4 392.2 7865. -463. -1367. 514.2 -1038.7 -40.4 99 -6528. 4618. 1091.4 -392.2 -7865. 463. 1367. -514.2 1038.7 40.4 100 1367. 6396. 934.0 0.0 6393. -165. -1367. 596.7 718.5 0.0 100 -1367. 6396. 934.0 0.0 -6393. 165. 1367. -596.7 -718.5 0.0 110 1367. 11038. 6985.2 0.0 -10639. 2944. -1367. -3736.4 -5901.8 0.0 110 0. 11545. 6985.2 0.0 -9443. 6641. 0. 3736.4 5901.8 0.0 120 0. 9099. 1939.5 0.0 6713. -6143. 0. -540.4 -1862.7 0.0 120 0. 9099. 1939.5 0.0 -6713. 6143. 0. 540.4 1862.7 0.0 130 0. 8265. 1124.3 0.0 6074. -5605. 0. 146.9 -1114.6 0.0 130 0. 8265. 1124.3 0.0 -6074. 5605. 0. -146.9 1114.6 0.0 140 0. 1438. 1959.9 0.0 1422. 218. 0. 1677.1 1014.2 0.0 140 0. 1438. 1959.9 0.0 -1422. -218. 0. -1677.1 -1014.2 0.0 150 0. 1088. 1982.8 0.0 783. 755. 0. 1620.2 1143.1 0.0 150 0. 1088. 1982.8 0.0 -783. -755. 0. -1620.2 -1143.1 0.0 160 0. 10505. 8426.1 0.0 -10139. 2749. 0. -1883.6 -8212.9 0.0 160 0. 13207. 8426.1 0.0 -12975. 2464. 0. 1883.6 8212.9 0.0 170 0. 8998. 0.0 0.0 -8868. 1523. 0. 0.0 0.0 0.0

178

CAESAR II Ver.4.40 Page: 25 FORCE REPORT, Forces on Elements CASE 8 (OCC) L8=L2+L4 ELEMENT AXIAL SHEAR BENDING TORSION -----------Forces(N. )----------- ---------Moments(N.m. )------ NODES: FORCE FORCE MOMENT MOMENT FX FY FZ MX MY MZ 10 -12266. 3205. 5037.5 -20.6 -753. 3115. -12266. 4950.4 -932.7 -20.6 20 12266. 3205. 6531.8 20.6 753. -3115. 12266. -6508.0 556.3 20.6 30 -3115. 12289. 6508.1 556.3 -753. 3115. -12266. 6508.0 -556.3 -20.6 40 2949. 11564. 4925.6 -556.3 753. -2949. 11540. -4924.9 556.3 -79.5 40 -2949. 11564. 4925.6 556.3 -753. 2949. -11540. 4924.9 -556.3 79.5 50 2411. 10927. 3616.0 -556.3 753. -2411. 10901. -3612.1 556.3 -167.6 50 -2411. 10927. 3616.0 556.3 -753. 2411. -10901. 3612.1 -556.3 167.6 60 1874. 10290. 2387.8 -556.3 753. -1874. 10262. -2374.1 556.3 -255.7 60 -1874. 10290. 2387.8 556.3 -753. 1874. -10262. 2374.1 -556.3 255.7 69 7330. 5157. 18.9 -658.1 753. -1575. 8791. 16.2 472.3 -458.5 69 -7330. 5157. 18.9 658.1 -753. 1575. -8791. -16.2 -472.3 458.5 70 8182. 1482. 650.5 -542.5 753. -1277. 8182. 592.1 269.5 -542.5 70 -8182. 1482. 650.5 542.5 -753. 1277. -8182. -592.1 -269.5 542.5 78 8182. 1215. 1433.8 -542.5 753. 954. 8182. 230.3 -1415.2 -542.5 78 -8182. 1215. 1433.8 542.5 -753. -954. -8182. -230.3 1415.2 542.5 79 5887. 4982. 1130.3 -103.0 753. 1252. 7572. 525.2 -749.9 -670.9 79 -5887. 4982. 1130.3 103.0 -753. -1252. -7572. -525.2 749.9 670.9 80 753. 6295. 1448.4 676.2 753. 1550. 6101. 676.2 997.1 -1050.5 80 -753. 6295. 1448.4 -676.2 -753. -1550. -6101. -676.2 -997.1 1050.5 90 753. 3478. 4328.9 676.2 753. 2167. 2721. 676.2 3727.5 -2201.1 90 -649. 3438. 4328.9 -676.2 -649. 2102. -2721. -676.2 -3727.5 2201.1 98 649. 6140. 981.3 676.2 649. -489. -6120. 676.2 975.8 -103.6 98 -649. 6140. 981.3 -676.2 -649. 489. 6120. -676.2 -975.8 103.6 99 5826. 4913. 930.1 508.2 649. -190. -7591. 708.8 -787.9 -9.9 99 -5826. 4913. 930.1 -508.2 -649. 190. 7591. -708.8 787.9 9.9 100 8201. 657. 1647.9 0.0 649. 108. -8201. 717.7 -1483.4 0.0 100 -8201. 657. 1647.9 0.0 -649. -108. 8201. -717.7 1483.4 0.0 110 8201. 3281. 4499.3 0.0 649. 3217. -8201. -4466.9 539.5 0.0 110 0. 6901. 4499.3 0.0 190. 6899. 0. 4466.9 -539.5 0.0 120 0. 6403. 1225.6 0.0 -190. -6400. 0. -1142.1 444.4 0.0 120 0. 6403. 1225.6 0.0 190. 6400. 0. 1142.1 -444.4 0.0 130 0. 5866. 598.9 0.0 -190. -5863. 0. -424.8 422.2 0.0 130 0. 5866. 598.9 0.0 190. 5863. 0. 424.8 -422.2 0.0 140 0. 194. 1290.4 0.0 -190. -40. 0. 1251.5 314.3 0.0 140 0. 194. 1290.4 0.0 190. 40. 0. -1251.5 -314.3 0.0 150 0. 533. 1259.1 0.0 -190. 498. 0. 1224.7 292.0 0.0 150 0. 533. 1259.1 0.0 190. -498. 0. -1224.7 -292.0 0.0 160 0. 2499. 1766.6 0.0 -190. 2491. 0. -1764.4 -88.1 0.0 160 0. 2435. 1766.6 0.0 -22. 2435. 0. 1764.4 88.1 0.0 170 0. 1553. 0.0 0.0 22. 1552. 0. 0.0 0.0 0.0

179

CAESAR II Ver.4.40 Page: 26 FORCE REPORT, Forces on Elements CASE 9 (OPE) L9=L1-L3 ELEMENT AXIAL SHEAR BENDING TORSION -----------Forces(N. )----------- ---------Moments(N.m. )------ NODES: FORCE FORCE MOMENT MOMENT FX FY FZ MX MY MZ 10 -1847. 12968. 11195.9 7224.4 12783. 2185. -1847. 1236.5 11127.4 7224.4 20 1847. 12968. 5277.5 -7224.4 -12783. -2185. 1847. -2329.0 -4735.8 -7224.4 30 -2185. 12916. 7590.6 -4735.8 12783. 2185. -1847. 2329.0 4735.8 7224.4 40 2018. 12197. 5949.3 4735.8 -12057. -2018. 1847. -2083.4 -4735.8 -5572.6 40 -2018. 12197. 5949.3 -4735.8 12057. 2018. -1847. 2083.4 4735.8 5572.6 50 1481. 11566. 4595.7 4735.8 -11418. -1481. 1847. -1867.2 -4735.8 -4199.3 50 -1481. 11566. 4595.7 -4735.8 11418. 1481. -1847. 1867.2 4735.8 4199.3 60 943. 10936. 3337.7 4735.8 -10779. -943. 1847. -1651.1 -4735.8 -2900.7 60 -943. 10936. 3337.7 -4735.8 10779. 943. -1847. 1651.1 4735.8 2900.7 69 1762. 9095. 3006.9 2435.0 -9055. -645. 1847. -1236.8 -3659.8 -216.2 69 -1762. 9095. 3006.9 -2435.0 9055. 645. -1847. 1236.8 3659.8 216.2 70 1847. 7340. 1870.0 728.8 -7332. -347. 1847. -1162.0 -1465.1 728.8 70 -1847. 7340. 1870.0 -728.8 7332. 347. -1847. 1162.0 1465.1 -728.8 78 1847. 5240. 1385.6 728.8 4889. 1884. 1847. 558.0 1268.3 728.8 78 -1847. 5240. 1385.6 -728.8 -4889. -1884. -1847. -558.0 -1268.3 -728.8 79 5804. 3866. 430.2 1131.5 6360. 2182. 1847. 1103.6 -30.1 496.6 79 -5804. 3866. 430.2 -1131.5 -6360. -2182. -1847. -1103.6 30.1 -496.6 80 6970. 3093. 297.1 1358.4 6970. 2480. 1847. 1358.4 -265.4 -133.6 80 -6970. 3093. 297.1 -1358.4 -6970. -2480. -1847. -1358.4 265.4 133.6 90 6970. 3606. 2056.8 1358.4 6970. 3097. 1847. 1358.4 878.0 -1860.0 90 11008. 2673. 2056.8 -1358.4 11008. 1932. -1847. -1358.4 -878.0 1860.0 98 -11008. 1874. 3868.8 1358.4 -11008. -318. 1847. 1358.4 3868.6 -38.6 98 11008. 1874. 3868.8 -1358.4 11008. 318. -1847. -1358.4 -3868.6 38.6 99 -8659. 6046. 3329.2 963.7 -10398. -20. 1847. 1371.9 3182.7 9.1 99 8659. 6046. 3329.2 -963.7 10398. 20. -1847. -1371.9 -3182.7 -9.1 100 -1847. 8931. 1554.5 0.0 -8927. 278. 1847. 1334.9 796.5 0.0 100 1847. 8931. 1554.5 0.0 8927. -278. -1847. -1334.9 -796.5 0.0 110 -1847. 8784. 4408.4 0.0 8105. 3387. 1847. -4381.6 -485.4 0.0 110 0. 9946. 4408.4 0.0 7193. 6869. 0. 4381.6 485.4 0.0 120 0. 7778. 3564.5 0.0 -4463. -6370. 0. -1071.9 -3399.5 0.0 120 0. 7778. 3564.5 0.0 4463. 6370. 0. 1071.9 3399.5 0.0 130 0. 6974. 3900.8 0.0 -3824. -5833. 0. -358.0 -3884.3 0.0 130 0. 6974. 3900.8 0.0 3824. 5833. 0. 358.0 3884.3 0.0 140 0. 829. 4910.6 0.0 829. -10. 0. 1301.2 -4735.0 0.0 140 0. 829. 4910.6 0.0 -829. 10. 0. -1301.2 4735.0 0.0 150 0. 1559. 4773.0 0.0 1467. 528. 0. 1270.9 -4600.7 0.0 150 0. 1559. 4773.0 0.0 -1467. -528. 0. -1270.9 4600.7 0.0 160 0. 12643. 9424.9 0.0 12389. 2521. 0. -1778.3 9255.6 0.0 160 0. 13458. 9424.9 0.0 13235. 2438. 0. 1778.3 -9255.6 0.0 170 0. 8746. 0.0 0.0 8608. 1549. 0. 0.0 0.0 0.0

180

CAESAR II Ver.4.40 Page: 27 FORCE REPORT, Forces on Elements CASE 10 (OPE) L10=L1-L4 ELEMENT AXIAL SHEAR BENDING TORSION -----------Forces(N. )----------- ---------Moments(N.m. )------ NODES: FORCE FORCE MOMENT MOMENT FX FY FZ MX MY MZ 10 11786. 2639. 7200.2 895.8 1860. 1872. 11786. -6135.4 3768.2 895.8 20 -11786. 2639. 5923.5 -895.8 -1860. -1872. -11786. 5199.4 -2838.1 -895.8 30 -1872. 11932. 5276.0 -2838.1 1860. 1872. 11786. -5199.4 2838.1 895.8 40 1706. 11215. 3736.7 2838.1 -1860. -1706. -11060. 3680.1 -2838.1 -648.4 40 -1706. 11215. 3736.7 -2838.1 1860. 1706. 11060. -3680.1 2838.1 648.4 50 1168. 10586. 2461.4 2838.1 -1860. -1168. -10421. 2423.4 -2838.1 -430.7 50 -1168. 10586. 2461.4 -2838.1 1860. 1168. 10421. -2423.4 2838.1 430.7 60 630. 9958. 1259.7 2838.1 -1860. -630. -9782. 1241.5 -2838.1 -213.1 60 -630. 9958. 1259.7 -2838.1 1860. 630. 9782. -1241.5 2838.1 213.1 69 -5642. 6389. 2078.1 2063.8 -1860. -332. -8311. -1255.0 -2630.5 288.1 69 5642. 6389. 2078.1 -2063.8 1860. 332. 8311. 1255.0 2630.5 -288.1 70 -7702. 1861. 3066.3 495.7 -1860. -34. -7702. -2206.4 -2129.3 495.7 70 7702. 1861. 3066.3 -495.7 1860. 34. 7702. 2206.4 2129.3 -495.7 78 -7702. 2879. 2045.2 495.7 -1860. 2197. -7702. 213.8 2034.0 495.7 78 7702. 2879. 2045.2 -495.7 1860. -2197. 7702. -213.8 -2034.0 -495.7 79 -6330. 4462. 1774.7 758.2 -1860. 2495. -7092. 843.6 1720.6 228.6 79 6330. 4462. 1774.7 -758.2 1860. -2495. 7092. -843.6 -1720.6 -228.6 80 -1860. 6277. 536.1 1133.4 -1860. 2793. -5621. 1133.4 226.4 -486.0 80 1860. 6277. 536.1 -1133.4 1860. -2793. 5621. -1133.4 -226.4 486.0 90 -1860. 4081. 3264.8 1133.4 -1860. 3410. -2241. 1133.4 -2206.9 -2406.0 90 3182. 3144. 3264.8 -1133.4 3182. 2205. 2241. -1133.4 2206.9 2406.0 98 -3182. 6626. 1329.3 1133.4 -3182. -591. 6600. 1133.4 1321.7 -142.6 98 3182. 6626. 1329.3 -1133.4 3182. 591. -6600. -1133.4 -1321.7 142.6 99 -7957. 3469. 3043.8 847.6 -3182. -293. 8071. 1177.4 2931.9 -21.4 99 7957. 3469. 3043.8 -847.6 3182. 293. -8071. -1177.4 -2931.9 21.4 100 -8680. 3182. 3234.9 0.0 -3182. 5. 8680. 1213.9 2998.5 0.0 100 8680. 3182. 3234.9 0.0 3182. -5. -8680. -1213.9 -2998.5 0.0 110 -8680. 4452. 7830.1 0.0 -3182. 3114. 8680. -3651.2 -6926.7 0.0 110 0. 7047. 7830.1 0.0 -2440. 6611. 0. 3651.2 6926.7 0.0 120 0. 6582. 5726.0 0.0 2440. -6113. 0. -470.1 -5706.6 0.0 120 0. 6582. 5726.0 0.0 -2440. 6113. 0. 470.1 5706.6 0.0 130 0. 6086. 5425.4 0.0 2440. -5575. 0. 213.7 -5421.1 0.0 130 0. 6086. 5425.4 0.0 -2440. 5575. 0. -213.7 5421.1 0.0 140 0. 2453. 4389.0 0.0 2440. 248. 0. 1726.7 -4035.1 0.0 140 0. 2453. 4389.0 0.0 -2440. -248. 0. -1726.7 4035.1 0.0 150 0. 2563. 4103.2 0.0 2440. 785. 0. 1666.3 -3749.6 0.0 150 0. 2563. 4103.2 0.0 -2440. -785. 0. -1666.3 3749.6 0.0 160 0. 3698. 2209.0 0.0 2440. 2779. 0. -1897.6 1130.8 0.0 160 0. 2484. 2209.0 0.0 283. 2468. 0. 1897.6 -1130.8 0.0 170 0. 1545. 0.0 0.0 -283. 1519. 0. 0.0 0.0 0.0

181

CAESAR II Ver.4.40 Page: 28 FORCE REPORT, Forces on Elements CASE 11 (OCC) L11=L2-L3 ELEMENT AXIAL SHEAR BENDING TORSION -----------Forces(N. )----------- ---------Moments(N.m. )------ NODES: FORCE FORCE MOMENT MOMENT FX FY FZ MX MY MZ 10 -1418. 11921. 8132.0 6921.8 11623. 2650. -1418. -266.0 8127.6 6921.8 20 1418. 11921. 2546.7 -6921.8 -11623. -2650. 1418. -1058.9 -2316.1 -6921.8 30 -2650. 11709. 7002.3 -2316.1 11623. 2650. -1418. 1058.9 2316.1 6921.8 40 2483. 10988. 5493.6 2316.1 -10897. -2483. 1418. -870.3 -2316.1 -5424.2 40 -2483. 10988. 5493.6 -2316.1 10897. 2483. -1418. 870.3 2316.1 5424.2 50 1946. 10355. 4245.5 2316.1 -10258. -1946. 1418. -704.4 -2316.1 -4186.7 50 -1946. 10355. 4245.5 -2316.1 10258. 1946. -1418. 704.4 2316.1 4186.7 60 1408. 9723. 3071.5 2316.1 -9619. -1408. 1418. -538.6 -2316.1 -3023.9 60 -1408. 9723. 3071.5 -2316.1 9619. 1408. -1418. 538.6 2316.1 3023.9 69 1787. 7898. 1458.9 507.5 -7895. -1110. 1418. -291.8 -1369.6 -651.9 69 -1787. 7898. 1458.9 -507.5 7895. 1110. -1418. 291.8 1369.6 651.9 70 1418. 6225. 644.2 163.6 -6172. -811. 1418. -390.2 512.5 163.6 70 -1418. 6225. 644.2 -163.6 6172. 811. -1418. 390.2 -512.5 -163.6 78 1418. 6214. 711.1 163.6 6049. 1419. 1418. 289.9 649.3 163.6 78 -1418. 6214. 711.1 -163.6 -6049. -1419. -1418. -289.9 -649.3 -163.6 79 6320. 4645. 1132.9 490.4 7521. 1718. 1418. 710.3 -1009.6 -16.7 79 -6320. 4645. 1132.9 -490.4 -7521. -1718. -1418. -710.3 1009.6 16.7 80 8130. 2464. 1578.7 913.3 8130. 2016. 1418. 913.3 -1490.0 -521.8 80 -8130. 2464. 1578.7 -913.3 -8130. -2016. -1418. -913.3 1490.0 521.8 90 8130. 2990. 2053.9 913.3 8130. 2633. 1418. 913.3 -612.5 -1960.5 90 8592. 2437. 2053.9 -913.3 8592. 1982. -1418. -913.3 612.5 1960.5 98 -8592. 1465. 1683.7 913.3 -8592. -368. 1418. 913.3 1682.7 -57.8 98 8592. 1465. 1683.7 -913.3 8592. 368. -1418. -913.3 -1682.7 57.8 99 -6647. 4643. 1327.5 656.8 -7982. -70. 1418. 932.4 1150.7 3.5 99 6647. 4643. 1327.5 -656.8 7982. 70. -1418. -932.4 -1150.7 -3.5 100 -1418. 6515. 1107.4 0.0 -6511. 228. 1418. 908.9 -632.6 0.0 100 1418. 6515. 1107.4 0.0 6511. -228. -1418. -908.9 632.6 0.0 110 -1418. 11037. 7294.9 0.0 10521. 3337. 1418. -4650.9 5620.0 0.0 110 0. 11653. 7294.9 0.0 9344. 6963. 0. 4650.9 -5620.0 0.0 120 0. 9249. 2081.4 0.0 -6614. -6465. 0. -1293.7 1630.5 0.0 120 0. 9249. 2081.4 0.0 6614. 6465. 0. 1293.7 -1630.5 0.0 130 0. 8416. 1059.6 0.0 -5975. -5927. 0. -568.8 894.1 0.0 130 0. 8416. 1059.6 0.0 5975. 5927. 0. 568.8 -894.1 0.0 140 0. 1326. 1642.6 0.0 -1322. -105. 0. 1144.4 -1178.4 0.0 140 0. 1326. 1642.6 0.0 1322. 105. 0. -1144.4 1178.4 0.0 150 0. 809. 1716.1 0.0 -683. 433. 0. 1125.1 -1295.7 0.0 150 0. 809. 1716.1 0.0 683. -433. 0. -1125.1 1295.7 0.0 160 0. 10522. 8439.0 0.0 10238. 2426. 0. -1734.3 8258.8 0.0 160 0. 13211. 8439.0 0.0 12986. 2427. 0. 1734.3 -8258.8 0.0 170 0. 8993. 0.0 0.0 8857. 1560. 0. 0.0 0.0 0.0

182

CAESAR II Ver.4.40 Page: 29 FORCE REPORT, Forces on Elements CASE 12 (OCC) L12=L2-L4 ELEMENT AXIAL SHEAR BENDING TORSION -----------Forces(N. )----------- ---------Moments(N.m. )------ NODES: FORCE FORCE MOMENT MOMENT FX FY FZ MX MY MZ 10 12216. 2439. 7676.5 593.1 700. 2337. 12216. -7637.9 768.4 593.1 20 -12216. 2439. 6483.1 -593.1 -700. -2337. -12216. 6469.5 -418.4 -593.1 30 -2337. 12236. 6496.7 -418.4 700. 2337. 12216. -6469.5 418.4 593.1 40 2170. 11511. 4918.6 418.4 -700. -2170. -11490. 4893.1 -418.4 -500.0 40 -2170. 11511. 4918.6 -418.4 700. 2170. 11490. -4893.1 418.4 500.0 50 1633. 10873. 3610.5 418.4 -700. -1633. -10851. 3586.2 -418.4 -418.1 50 -1633. 10873. 3610.5 -418.4 700. 1633. 10851. -3586.2 418.4 418.1 60 1095. 10236. 2377.9 418.4 -700. -1095. -10212. 2354.0 -418.4 -336.2 60 -1095. 10236. 2377.9 -418.4 700. 1095. 10212. -2354.0 418.4 336.2 69 -5617. 6780. 463.8 136.2 -700. -797. -8741. -310.0 -340.3 -147.6 69 5617. 6780. 463.8 -136.2 700. 797. 8741. 310.0 340.3 147.6 70 -8131. 859. 1442.6 -69.5 -700. -499. -8131. -1434.6 -151.7 -69.5 70 8131. 859. 1442.6 69.5 700. 499. 8131. 1434.6 151.7 69.5 78 -8131. 1868. 1416.1 -69.5 -700. 1732. -8131. -54.3 1415.0 -69.5 78 8131. 1868. 1416.1 69.5 700. -1732. 8131. 54.3 -1415.0 69.5 79 -5814. 5234. 905.2 117.1 -700. 2030. -7522. 450.3 741.1 -284.7 79 5814. 5234. 905.2 -117.1 700. -2030. 7522. -450.3 -741.1 284.7 80 -700. 6483. 1326.8 688.2 -700. 2329. -6051. 688.2 -998.2 -874.1 80 700. 6483. 1326.8 -688.2 700. -2329. 6051. -688.2 998.2 874.1 90 -700. 3976. 4466.9 688.2 -700. 2946. -2670. 688.2 -3697.4 -2506.5 90 766. 3495. 4466.9 -688.2 766. 2255. 2670. -688.2 3697.4 2506.5 98 -766. 6204. 879.2 688.2 -766. -641. 6170. 688.2 -864.2 -161.8 98 766. 6204. 879.2 -688.2 766. 641. -6170. -688.2 864.2 161.8 99 -5945. 4874. 1030.8 540.8 -766. -343. 7642. 737.9 899.9 -27.0 99 5945. 4874. 1030.8 -540.8 766. 343. -7642. -737.9 -899.9 27.0 100 -8251. 768. 1756.0 0.0 -766. -45. 8251. 787.9 1569.3 0.0 100 8251. 768. 1756.0 0.0 766. 45. -8251. -787.9 -1569.3 0.0 110 -8251. 3158. 4005.5 0.0 -766. 3064. 8251. -3920.4 -821.3 0.0 110 0. 6712. 4005.5 0.0 -289. 6706. 0. 3920.4 821.3 0.0 120 0. 6214. 967.8 0.0 289. -6208. 0. -691.9 -676.6 0.0 120 0. 6214. 967.8 0.0 -289. 6208. 0. 691.9 676.6 0.0 130 0. 5678. 642.8 0.0 289. -5670. 0. 2.9 -642.8 0.0 130 0. 5678. 642.8 0.0 -289. 5670. 0. -2.9 642.8 0.0 140 0. 327. 1641.2 0.0 289. 153. 0. 1569.9 -478.4 0.0 140 0. 327. 1641.2 0.0 -289. -153. 0. -1569.9 478.4 0.0 150 0. 749. 1584.2 0.0 289. 690. 0. 1520.6 -444.6 0.0 150 0. 749. 1584.2 0.0 -289. -690. 0. -1520.6 444.6 0.0 160 0. 2699. 1858.4 0.0 289. 2684. 0. -1853.6 134.1 0.0 160 0. 2457. 1858.4 0.0 34. 2457. 0. 1853.6 -134.1 0.0 170 0. 1530. 0.0 0.0 -34. 1530. 0. 0.0 0.0 0.0

183

CAESAR II Ver.4.40 Page: i LISTING OF STATIC LOAD CASES FOR THIS ANALYSIS CASE --------TITLE--------- CASE 1 (OPE) W+T1+P1 CASE 2 (SUS) W+P1 CASE 3 (SUS) WIN1 CASE 4 (SUS) WIN3 CASE 5 (OPE) L5=L1+L3 CASE 6 (OPE) L6=L1+L4 CASE 7 (OCC) L7=L2+L3 CASE 8 (OCC) L8=L2+L4 CASE 9 (OPE) L9=L1-L3 CASE 10 (OPE) L10=L1-L4 CASE 11 (OCC) L11=L2-L3 CASE 12 (OCC) L12=L2-L4 CASE 13 (EXP) L13=L1-L2 TABLE OF CONTENTS -- STATIC OUTPUT -----------------REPORT----------------- ----------LOAD CASE---------- -PAGE- INPUT LISTING - N/A - 1 RESTRAINT SUMMARY VARIOUS 6 STRESS SUMMARY CASE 2 (SUS) W+P1 8 STRESS REPORT, Stresses on Elements CASE 2 (SUS) W+P1 9 STRESS SUMMARY CASE 7 (OCC) L7=L2+L3 10 STRESS REPORT, Stresses on Elements CASE 7 (OCC) L7=L2+L3 11 STRESS SUMMARY CASE 8 (OCC) L8=L2+L4 12 STRESS REPORT, Stresses on Elements CASE 8 (OCC) L8=L2+L4 13 STRESS SUMMARY CASE 11 (OCC) L11=L2-L3 14 STRESS REPORT, Stresses on Elements CASE 11 (OCC) L11=L2-L3 15 STRESS SUMMARY CASE 12 (OCC) L12=L2-L4 16 STRESS REPORT, Stresses on Elements CASE 12 (OCC) L12=L2-L4 17 STRESS SUMMARY CASE 13 (EXP) L13=L1-L2 18 STRESS REPORT, Stresses on Elements CASE 13 (EXP) L13=L1-L2 19 FORCE/STRESS REPORT, Forces on Elements CASE 1 (OPE) W+T1+P1 20 FORCE/STRESS REPORT, Forces on Elements CASE 2 (SUS) W+P1 21 FORCE/STRESS REPORT, Forces on Elements CASE 5 (OPE) L5=L1+L3 22 FORCE/STRESS REPORT, Forces on Elements CASE 6 (OPE) L6=L1+L4 23 FORCE/STRESS REPORT, Forces on Elements CASE 7 (OCC) L7=L2+L3 24 FORCE/STRESS REPORT, Forces on Elements CASE 8 (OCC) L8=L2+L4 25 FORCE/STRESS REPORT, Forces on Elements CASE 9 (OPE) L9=L1-L3 26 FORCE/STRESS REPORT, Forces on Elements CASE 10 (OPE) L10=L1-L4 27 FORCE/STRESS REPORT, Forces on Elements CASE 11 (OCC) L11=L2-L3 28 FORCE/STRESS REPORT, Forces on Elements CASE 12 (OCC) L12=L2-L4 29

184

EXAMPLE 2 Introduction Example 2 is the same process segment as in Example 1, i.e. the produced hydrocarbon (HC) liquid pipe downstream from a 2nd stage separator, but now subject to accidental jet fire. Only the stress conditions in the pipe are of the concern, as this Technical Note does not address pressure vessels. The process segment includes the pipework and equipment from the emergency shutdown valve (ESDV) upstream the separator, the separator itself and the produced HC liquid pipe downstream the separator to the ESDV on the pipe. A simplistic criticality assessment was carried out, which resulted in the conclusion that a multi-physics analysis was to be carried out to determine the pressure variation in the pipe affected by the fire, which may be followed by a pipe stress analysis. The multi-physics analysis of the time-variation of the internal pressure was carried out using computer program VessFire [1, 2] and the pipe stress analysis may be performed in the application of computer program Caesar II [3]. The VessFire analysis showed a pipe failure and, following from that, the wall thickness of the pipe was increased. A second VessFire analysis confirmed that the pipe with the increased wall thickness survived and maintained containment throughout the fire. Two combinations of the pipe temperature and internal pressure, which should give the worst conditions, were determined from the VessFire results and these may be transferred to a Caesar II analysis, where a linear elastic pipe stress analysis may be performed. Revising to 20mm wall thickness does increase the static operating loads, but they are still acceptable (original loads were very low). In this example it would be OK. It should be recognised that it is very useful, and inherently safe practice, to specify the potential for thicker wall very early in the project life, in which case the vessel can be designed to accommodate higher loads. Process Segment The pipe is a part of the 2nd stage separator process segment, where the process segment extends from the ESDV located immediately upstream the separator, the separator itself, and the produced HC liquid pipe to the ESDV downstream the separator. The process segment is equipped with a pressure relief valve (PSV) and an emergency depressurisation (blowdown) valve (BDV). Although the objective of the PSV is primarily to protect the process segment from overpressure in the case of upset process conditions, the PSV will also relieve pressure in accidental conditions, where the segment internal pressure may rise due to the heating-up of the segment inventory. The objective of the blowdown system is to protect the process segment in the accidental case of a fire. The whole process segment is in a naturally ventilated module enclosure, which forms a fire area. Fire Scenario The pipe is exposed to a hydrocarbon jet fire that affects the whole segment. The source of the fire is considered to be from the outside of the segment. As the heat flux from the fire is reflected from the walls of the enclosure back into the flame, the fire loads are as follows:

185

• a point load heat flux from the engulfing jet flame onto the pipe of 350kW/m2; and • a background heat flux due to the heat flux reflected from the enclosure walls of the intensity of

100kW/m2, which affects the whole segment. The fire duration is 15 minutes. The total volume of hydrocarbon in the segment affected by the background heat flux is 12.68m3, which is comprised of: • 0.62m3 of liquid HC in the pipe; • 12.06m3 of HC liquid and gas, and water in the separator. The HC volume in the pipe between the separator and the ESDV upstream of the separator is considered negligible. There is 50% HC liquid and water in the separator, and 50% HC gas. The HC liquid pipe downstream the separator contains HC liquid only. The stream composition in the segment is shown in Table A1. Emergency Response The process segment emergency response data that concerns the analysis is as follows: • The ESDVs are fully shut at 60s after the start of the fire. • The BDV is fully open at 60s after the start of the fire. Criticality The likelihood of the fire scenario is assumed to be above 1x10-4. This is due to the large number of potential sources of accidental release of hydrocarbon in the segment. The fire originates from the outside of the segment and it may cause the loss of containment from the segment. This would result in a combination of an additional gas jet fire and a prolonged pool fire caused by the large volume of hydrocarbon in the segment, which would leak out and ignite. Such consequences would be very severe as the event may escalate to a neighbouring fire area beyond the module enclosure. Following from these considerations, the potential escalation to the neighbouring module must be prevented. The jet fire must not cause the loss of containment in the process segment. The criticality is considered to be level 1. Methodology Based on the criticality level 1 of the fire scenario, the analysis has been carried out in the following three steps: Step 1: A multi-physics analysis of the time-dependent pressure in the pipe, which will give the pressure time variation and a simplified stress distribution in the pipe. The multi-physics analysis considers the pipe as a straight cylinder. Step 2: Based on the results from the multi-physics analysis, the selection of the critical combination of pressure and temperature in the pipe and the transfer of this data to carry out any pipe stress analysis. Step 3: The pipe stress analysis, which will take into account the routing of the pipe, pipe supports, valves, etc. 186

Multi-Physics Analysis The multi-physics analysis in Step 1 was carried out using computer program VessFire. The process segment was modelled as an infinite pipe, using the data summarised in Table A1.

Table A1. Data for multi-physics analysis of pressure loading in the liquid outlet pipe from 2nd stage

separator.

Description Data No Pipe Typically, a liquid line from 2nd stage separator 1 Pipe inner diameter 273.05mm 2 Wall thickness 9.271mm 3 Corrosion allowance 3mm (in addition to wall thickness) 4 Volume of the segment 12.68m3, whereof 50% gas (in the separator) 5 Pipe operating pressure 15barg 6 Pipe design pressure 20barg 7 Segment high-high pressure 20barg 8 Pipe hydro test pressure 25barg 9 Pipe operating temperature 40oC

10 Pipe design temperature -10oC/+100oC 11 Pipe material A333 Grade 6 (carbon steel) 12 Density 7850kg/m3

13 Specific heat 440J/kgK (at 20oC, varying with material temperature) 14 Thermal conductivity 53W/mK (at 20oC, varying with material temperature) 15 Yield stress 241Mpa 16 Ultimate Tensile Strength 413.7MPa

Pipe Inventory 17 Inventory characteristics Separated hydrocarbon liquid 18 Density 815kg/m3

Fluid Composition Mol(%) 19 Nitrogen (N2) 0.0002 20 Carbon dioxide (CO2) 0.0006 21 Hydrogen sulphide (H2S) 0 22 Methane (C1) 0.2103 23 Ethane (C2) 0.0183 24 Propane (C3) 0.0351 25 I-Butane (i-C4) 0.0090 26 N-Butane (n-C4) 0.0312 27 I-Pentane (i-C5) 0.0138 28 N-Pentane (n-C5) 0.0218 29 N-Hexane (C6) 0.0310 30 N-Heptane (C7) 0.0459 31 N-Octane (C8) 0.0529 32 N-Nonane (C9) 0.0327 33 N-Tridecane (C13) 0.1495 34 Water (H2O) 0.3477

Segment Isolation by ESD 35 ESD response time ESD valve fully shut at 60s after the start of the fire.

Segment Pressure Relief and Emergency Depressurisation

36 To where Atmosphere 37 BD response time BD valve fully opened at 60s after the start of the fire. 38 Blowdown orifice diameter 15mm 39 Backpressure 1.0bara

187

40 Pressure relief valve (PSV)? Yes 41 PSV orifice size 12mm 42 PSV opening pressure 20barg 43 PSV closing pressure 15barg 44 PSV operating profile square

Pipe Cooling None 45 Firewater equipment n/a 46 Applied firewater rate n/a

Pipe Passive Fire Protection Coating

None

47 Material n/a 48 Thickness n/a 49 Density n/a 50 Specific heat n/a 51 Thermal conductivity n/a

188

Table A1. Data for multi-physics analysis of pressure loading in the liquid outlet pipe from 2nd stage separator. (cont’d)

Description Data

No Pipe Thermal Insulation None 52 Material n/a 53 Thickness n/a 54 Density n/a 55 Specific heat n/a 56 Thermal conductivity n/a 57 Compressive strength n/a 58 Outer surface covered by n/a

Fire Scenario Data 59 Fire type Jet fire in an enclosed module with sufficient supply of air. 60 Flame heat release rate 350kW/m2 (maximum point load), 100kW/m2 (background load). 61 Fire duration Rising at time = 0 in step from 0 to 350kW/m2 (100kW/m2) and

remaining constant for 15mins. Reducing to 0 at 15mins. 62 Engulfment The pipe is fully engulfed around its circumference by the point

load. The segment is fully engulfed by the background load.

Results of the Multi-Physics Analysis The results of the multi-physics analysis in the application of VessFire are summarised in Figures A1 and A2. The analysis was carried out for the duration of 20 minutes, i.e. 5 minutes longer than the fire duration. The Figures also indicate the cooling effects of the pipe after the fire stopped at 15 minutes. Figures A1a through to A1h show results for the pipe diameter of 9.271mm. Due to the exposure to the fire the temperature of the pipe and pipe contents increases (Figures A1c, A1d and A1h). Figures A1e and A1f show the time history of the mass in the segment and release rate through the PSV and BDV. As it may be seen in Figure A1e the hydrocarbon gas mass suddenly rises with a corresponding decrease of the liquid hydrocarbon mass. The change occurs as the conditions in the process segment become supercritical. In the supercritical region the fluid can be described as neither liquid nor gas, but as the further depressurisation follows that of expanding gas, the segment inventory has been depicted as gas. Following the evaporation of the segment contents and the thermal expansion of the fluids, the pressure inside the pipe increases although the segment is depressurised through both the BDV and PSV (Figure A1a). This is reflected in Figure 1b, which shows the time variation of the applied stress and the material yield stress, where the latter reduces with the rising temperature of the pipe. The pipe was deemed to fail when the applied stress was equal to the material yield stress. This assumes elastic-perfectly plastic material where, at the yield stress, the pipe deformation increases infinitely with no increase of the load. In reality, steel exhibits residual strength above the first yield, which is caused by its hardening. However, there is a substantial uncertainty in the experimental data in the hardening region. Therefore, it was the material yield stress that was taken as the failure criterion. As it may be seen in Figure A1b the pipe fails at approximately 6.2 minutes after the start of the fire, at which point the applied stress becomes equal to the material yield stress. Following the indication of the pipe failure the pipe diameter was increased to 20mm and the VessFire analysis was re-computed. The results are shown in Figures A2a to A2h. The increase of the wall thickness resulted in the reduction of the rate of temperature rise, and the evaporation and thermal expansion rate of the

189

pipe contents. The peak pressure reduced from 13200kPa at 11.5min (Figure A1a) to 10800kPa at 13min (Figure A2a) after the start of the fire. With the 20mm wall thickness, the pipe exhibits a greater strength, the pressure is lower, the rate of temperature rise is lower, and the material yield strength reduction rate is also lower. As Figure A2b shows the applied stress remains below the material yield stress, and the pipe survives and maintains containment throughout the whole duration of the 15 minutes jet fire without fire protection coating. It should be noted, however, that the pipe flanges will require fire protection as they would lose their tightness within the first 10 minutes after the start of the fire. Pipe Stress Analysis As the VessFire analysis indicated no failure, the results that may give worst cases may be transferred to Caesar II for pipe stress analysis in order to investigate the effects of the fire-induced pressure and temperature of the pipe onto its global behaviour in respect to supports and stress risers. Based on Figures A2a and A2c, the worst cases are represented by the combinations of pressure and temperature at 13 (for highest pressure) and 15 minutes (for highest temperature) after the start of the fire. The following pipe data may be applied in Caesar II: Case 1: Wall thickness 20mm; Internal pressure 10800kPa; Temperature 340oC. Case 2: Wall thickness 20mm; Internal pressure 9800kPa; Temperature 440oC. References 1. VESSFIRE, An Analysis System for Depressurisation of Process Segments and Process Equipment

Exposed and Not Exposed to Fire, User’s Manual, Petrell AS, 2002. 2. VESSFIRE, An Analysis System for Depressurisation of Process Segments and Process Equipment

Exposed and Not Exposed to Fire, Technical Reference Manual, Petrell AS, 2003. 3. CAESAR II, (HOLD – Paul Rattenbury, please write here the full title of the Caesar II User’s Manual).

190

Figure A1a

Time History of Internal Pressure in the Pipe

0

2,000

4,000

6,000

8,000

10,000

12,000

14,000

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Time [min]

Pre

ssur

e [k

Pa]

Pressure in pipe

Figure A1b

Time History of Material Yield Stress and Applied Stress in the Pipe Wall

0

50

100

150

200

250

300

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Time [min]

Stre

ss [N

/mm

2 ]

Yield stress Calculated stress in pipe wall

191

Figure A1c

Average Pipe Temperature

0.00

100.00

200.00

300.00

400.00

500.00

600.00

700.00

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Time [min]

Tem

pera

ture

[°C

]

Pipe temperature BDV-line

Figure A1d

Time History of the Temperature of Gas and Oil in the Segment

0.00

100.00

200.00

300.00

400.00

500.00

600.00

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Time [min]

Tem

pera

ture

[°C

]

Temperature gas Temperature oil

192

Figure A1e

Time History of Mass in the Segment

0

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

4,500

5,000

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Time [min]

Mas

s liq

uid

[kg]

0

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

Mas

s ga

s [k

g]

Mass oil Sum mass Mass water Mass gas Mass steam

Figure A1f

Release Rate

0.000

1.000

2.000

3.000

4.000

5.000

6.000

7.000

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Time [min]

Rel

ease

rate

[kg/

s]

Release rate

193

Figure A1g

Heat Flux Absorbed by the Pipe Wall

0

10,000

20,000

30,000

40,000

50,000

60,000

70,000

80,000

90,000

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Time [min]

Hea

t flu

x [W

/m2 ]

Sum heat flux Radiation flux Convective flux

Figure A1h

Time History of Temperature Distribution Through the Pipe Wall(T1 at the inner surface of the pipe, T12 at the outer surface of the pipe, the remaining

between the two locations)

0.00

100.00

200.00

300.00

400.00

500.00

600.00

700.00

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Time [min]

Tem

pera

ture

[°C

]

Tmin-1

Tmin-2

Tmin-3

Tmin-4

Tmin-5

Tmin-6

Tmin-7

Tmin-8

Tmin-9

Tmin-10

Tmin-11

Tmin-12

194

Figure A2a

Time History of Internal Pressure in the Pipe

0

2,000

4,000

6,000

8,000

10,000

12,000

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Time [min]

Pre

ssur

e [k

Pa]

Pressure in pipe

Figure A2b

Time History of Material Yield Stress and Applied Stress in the Pipe Wall

0

50

100

150

200

250

300

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Time [min]

Stre

ss [N

/mm

2 ]

Yield stress Calculated stress in pipe wall

195

Figure A2c

Average Pipe Temperature

0.00

50.00

100.00

150.00

200.00

250.00

300.00

350.00

400.00

450.00

500.00

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Time [min]

Tem

pera

ture

[°C

]

Pipe temperature BDV-line

Figure A2d

Time History of the Temperature of Gas and Oil in the Segment

0.00

50.00

100.00

150.00

200.00

250.00

300.00

350.00

400.00

450.00

500.00

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Time [min]

Tem

pera

ture

[°C

]

Temperature gas Temperature oil

196

Figure A2e

Time History of Mass in the Segment

0

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

4,500

5,000

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Time [min]

Mas

s liq

uid

[kg]

0

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

Mas

s ga

s [k

g]

Mass oil Sum Mass Mass water Mass gas Mass steam

Figure A2f

Release Rate

0.000

1.000

2.000

3.000

4.000

5.000

6.000

7.000

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Time [min]

Rel

ease

rate

[kg/

s]

Release rate

197

Figure A2g

Heat Flux Absorbed by the Pipe Wall

0

10,000

20,000

30,000

40,000

50,000

60,000

70,000

80,000

90,000

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Time [min]

Hea

t flu

x [W

/m2 ]

Sum heat flux Radiation flux Convective flux

Figure A2h

Time History of Temperature Distribution Through the Pipe Wall(T1 at the inner surface of the pipe, T12 at the outer surface of the pipe, the remaining

between the two locations)

0.00

50.00

100.00

150.00

200.00

250.00

300.00

350.00

400.00

450.00

500.00

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Time [min]

Tem

pera

ture

[°C

]

Tmin-1

Tmin-2

Tmin-3

Tmin-4

Tmin-5

Tmin-6

Tmin-7

Tmin-8

Tmin-9

Tmin-10

Tmin-11

Tmin-12

198

fabig technical note 8

Documents

explosions technical

fabig members

fabig activities

fabig deliverables

forewordthis technical

written permission of

fabig project manager

design of offshore piping