190 views

Uploaded by Muhammad Habib

DNV RP C204 Design Against Accidental Loads, ship platform collision

- Theory of Local Buckling
- DNV-RP-C203
- oto01063[1]
- HAZID Guideword Checklist
- Structural Reviewer
- Dnv Dal Spec.
- Offshore Manual
- biomechanics of the mandible.pdf
- Accidental
- Stress Strain& Stress Concentration Factor
- Tcm Strength
- 2001Artigo_Wang-Hsu_Nonlinear Finite Element Analysis of Concrete Structures Using New Constitutive Models
- EG-120
- Circular Concrete-Filled Tubular Columns- State of the Art Oriented to the Vulnerability Assessment
- Wp1
- Deformations of Reinforced Concrete Members at Yielding and Ultimate
- [Dr._Mirvat_Bulbul]_ENCE335_Fundamentals_of_Reinfo(bookzz.org).pdf
- Civil_Reinforced Concrete Structures
- deshpande.pdf
- 6cb5da3513bd26085ee3fad631ebb37a-original.pdf

You are on page 1of 53

OCTOBER 2010

FOREWORD

DET NORSKE VERITAS (DNV) is an autonomous and independent foundation with the objectives of safeguarding life, property and the environment, at sea and onshore. DNV undertakes classification, certification, and other verification and consultancy services relating to quality of ships, offshore units and installations, and onshore industries worldwide, and carries out research in relation to these functions. DNV service documents consist of amongst other the following types of documents: Service Specifications. Procedual requirements. Standards. Technical requirements. Recommended Practices. Guidance. The Standards and Recommended Practices are offered within the following areas: A) Qualification, Quality and Safety Methodology B) Materials Technology C) Structures D) Systems E) Special Facilities F) Pipelines and Risers G) Asset Operation H) Marine Operations J) Cleaner Energy

O) Subsea Systems

The electronic pdf version of this document found through http://www.dnv.com is the officially binding version Det Norske Veritas Any comments may be sent by e-mail to rules@dnv.com For subscription orders or information about subscription terms, please use distribution@dnv.com Computer Typesetting (Adobe Frame Maker) by Det Norske Veritas

If any person suffers loss or damage which is proved to have been caused by any negligent act or omission of Det Norske Veritas, then Det Norske Veritas shall pay compensation to such person for his proved direct loss or damage. However, the compensation shall not exceed an amount equal to ten times the fee charged for the service in question, provided that the maximum compensation shall never exceed USD 2 million. In this provision "Det Norske Veritas" shall mean the Foundation Det Norske Veritas as well as all its subsidiaries, directors, officers, employees, agents and any other acting on behalf of Det Norske Veritas.

CHANGES

General Main changes

As of October 2010 all DNV service documents are primarily published electronically. In order to ensure a practical transition from the print scheme to the electronic scheme, all documents having incorporated amendments and corrections more recent than the date of the latest printed issue, have been given the date October 2010. An overview of DNV service documents, their update status and historical amendments and corrections may be found through http://www.dnv.com/resources/rules_standards/.

Since the previous edition (November 2004), this document has been amended, most recently in April 2005. All changes have been incorporated and a new date (October 2010) has been given as explained under General.

CONTENTS

1. 1.1 1.2 1.3 1.4 1.5 1.6 2. 2.1 2.2 2.3 2.4 2.5 3. 3.1 3.2 3.3 GENERAL .............................................................. 7 Introduction .............................................................7 Application ...............................................................7 Objectives .................................................................7 Normative references ..............................................7

4.4.1 4.4.2 4.4.3 Stiffened plates subjected to drill collar impact ............... 21 Stiffeners/girders .............................................................. 21 Dropped object ................................................................. 21 Pipes on plated structures ................................................. 21 Blunt objects ..................................................................... 21

4.5

4.5.1 4.5.2

1.4.1 1.4.2

Definitions ................................................................7 Symbols.....................................................................8 DESIGN PHILOSOPHY ....................................... 9 General .....................................................................9 Safety format............................................................9 Accidental loads .......................................................9 Acceptance criteria..................................................9 Analysis considerations .........................................10 SHIP COLLISIONS............................................. 10 General ...................................................................10 Design principles....................................................10 Collision mechanics ...............................................11 Dissipation of strain energy ..................................11 Ship collision forces ...............................................11

FIRE ...................................................................... 21 General................................................................... 21 General calculation methods................................ 22 Material modelling................................................ 22 Equivalent imperfections...................................... 22 Empirical correction factor.................................. 22 Local cross sectional buckling.............................. 22 Ductility limits ....................................................... 22

General.............................................................................. 22 Beams in bending ............................................................. 23 Beams in tension............................................................... 23

5.8 6. 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9

6.9.1 6.9.2 6.9.3

Capacity of connections ........................................ 23 EXPLOSIONS ...................................................... 23 General................................................................... 23 Classification of response ..................................... 23 Recommended analysis models for stiffened panels...................................................................... 23 SDOF system analogy ........................................... 25 Dynamic response charts for SDOF system ....... 26 MDOF analysis...................................................... 27 Classification of resistance properties ................ 27

Cross-sectional behaviour................................................. 27

3.3.1 3.3.2

3.4 3.5

Recommended force-deformation relationships............... 11 Force contact area for strength design of large diameter columns............................................................................. 13 Energy dissipation is ship bow ......................................... 13

3.6 3.7

3.7.1 3.7.2 3.7.3 3.7.4

Force-deformation relationships for denting of tubular members ...................................................14 Force-deformation relationships for beams........14

6.7.1

General.............................................................................. 14 Plastic force-deformation relationships including elastic, axial flexibility.................................................................. 14 Support capacity smaller than plastic bending moment of the beam............................................................................ 16 Bending capacity of dented tubular members .................. 16

Idealisation of resistance curves .......................... 28 Resistance curves and transformation factors for plates ................................................................ 28

Elastic - rigid plastic relationships.................................... 28 Axial restraint ................................................................... 29 Tensile fracture of yield hinges ........................................ 29

6.10

6.10.1 6.10.2 6.10.3 6.10.4 6.10.5 6.10.6 6.10.7

General.............................................................................. 17 Local buckling ................................................................. 17 Tensile fracture ................................................................. 18 Tensile fracture in yield hinges......................................... 18

Beams with no- or full axial restraint ............................... 29 Beams with partial end restraint. ...................................... 32 Beams with partial end restraint - support capacity smaller than plastic bending moment of member............. 34 Effective flange................................................................. 34 Strength of adjacent structure ........................................... 34 Strength of connections .................................................... 34 Ductility limits.................................................................. 34

3.11

7. 8. 9. 9.1

REFERENCES..................................................... 35 COMMENTARY ................................................. 35 EXAMPLES ......................................................... 43 Design against ship collisions ............................... 43 Design against explosions ..................................... 44

Energy dissipation in floating production vessels......................................................................19 Global integrity during impact ............................19 DROPPED OBJECTS ......................................... 19 General ...................................................................19 Impact velocity.......................................................20 Dissipation of strain energy ..................................21 Resistance/energy dissipation ...............................21

9.1.1

9.2

Geometry .......................................................................... 44 Calculation of dynamic response of plate: ....................... 44 Calculation of dynamic response of stiffened plate.......... 44 Plates................................................................................. 44 Calculation of resistance curve for stiffened plate ........... 45 Calculation of resistance curve for girder......................... 46

9.3

9.4

9.4.1 9.4.2 9.4.3

Plating ...............................................................................46 Stiffener: ...........................................................................46 Girder: ...............................................................................47

9.5

Geometry, material and loads ...........................................47 Cross sectional of properties for the girder.......................48 Mass ..................................................................................51 Natural period ...................................................................51 Ductility ratio ....................................................................52 Maximum blast pressure capacity.....................................52

1. General

1.1 Introduction

This Recommended Practice deals with design to maintain the load-bearing function of the structures during accidental events. The overall goal of the design against accidental loads is to achieve a system where the main safety functions of the installation are not impaired. The Recommended Practice has been developed for general world-wide application. Governmental legislation may include requirements in excess of the provisions of this Recommended Practice depending on type, location and intended service of the unit/installation. The Design Accidental Loads and associated performance criteria are given in DNV-OS-A101. The Accidental Loads in this standard are prescriptive loads. This Recommended Practice may also be used in cases where the Design Accidental Loads are determined by a formal safety assessment (see DNV-OS-A101, Appendix C) or Quantified Risk Assessment (QRA). The following main subjects are covered: Design philosophy Ship Collisions Dropped Objects Fire Explosions.

DNV-OS-C301

1.4.2 DNV Recommended Practices (RP) The latest revision of the following documents applies: DNV-RP-C201 DNV-RP-C202 Buckling Strength of Plated Structures Buckling Strength of Shells

1.5 Definitions

Load-bearing structure: That part of the facility whose main function is to transfer loads. Accidental Event: An undesired incident or condition which, in combination with other conditions (e.g.: weather conditions, failure of safety barrier, etc.), determines the accidental effects. Accidental Effect: The result of an accidental event, expressed in terms of heat flux, impact force and energy, acceleration, etc. which is the basis for the safety evaluations. Design Accidental Event (DAE): An accidental event, which results in effects that, the platform should be designed to sustain. Acceptance criteria: Functional requirements, which are concerned with the platforms' resistance to accidental effects. This should be in accordance with the authority's definition of acceptable safety levels. Active protection: Operational loads and mechanical equipment which are brought into operation when an accident is threatening or after the accident has occurred, in order to limit the probability of the accident and the effects thereof, respectively. Some examples are safety valves, shut down systems, water drenching systems, working procedures, drills for coping with accidents, etc. Passive protection: Protection against damage by means of distance, location, strength and durability of structural elements, insulation, etc. Event control: Implementation of measures for reducing the probability and consequence of accidental events, such as changes and improvements in equipment, working procedures, active protection devices, arrangement of the platform, personnel training, etc. Indirect design: Implementation of measures for improving structural ductility and resistance without numerical calculations and determination of specific accidental effects. Direct design: Determination of structural resistance, dimensions, etc. on basis of specific design accidental effects. Load: Any action causing load effect in the structure. Characteristic load: Reference value of a load to be used in determination of load effects when using the partial coefficient method or the allowable stress method. Load effect: Effect of a single load or combination of loads on the structure, such as stress, stress resultant (internal force and moment), deformation, displacement, motion, etc. Resistance: Capability of a structure or part of a structure to resist load effect. Characteristic resistance: The nominal capacity that may be used for determination of design resistance of a structure or structural element. The characteristic value of resistance is to be based on a defined percentile of the test results. Design life: The time period from commencement of construction until condemnation of the structure. Limit state: A state where a criterion governing the load-carrying ability or use of the structure is reached.

1.2 Application

The Recommended Practice is applicable to all types of floating and fixed offshore structures made of steel. The methods described are relevant for both substructures and topside structures. The document is limited to load-carrying structures and does not cover pressurised equipment.

1.3 Objectives

The objective with this Recommended Practice is to provide recommendations for design of structures exposed to accidental events.

The following standards include requirements which, through reference in the text constitute provisions of this Recommended Practice. Latest issue of the references shall be used unless otherwise agreed. Other recognised standards may be used provided it can be demonstrated that these meet or exceed the requirements of the standards referenced below. Any deviations, exceptions and modifications to the codes and standards shall be documented and agreed between the supplier, purchaser and verifier, as applicable. 1.4.1 DNV Offshore Standards (OS) The latest revision of the following documents applies: DNV-OS-A101 DNV-OS-C101 DNV-OS-C102 DNV-OS-C103 DNV-OS-C104 DNV-OS-C105 DNV-OS-C106 Safety Principles and Arrangements Design of Offshore Steel Structures, General (LRFD Method) Structural Design of Offshore Ships Structural Design of Column Stabilised Units (LRFD) Structural Design of Self-Elevating Units (LRFD) Structural Design of TLPs (LRFD) Structural Design of Deep Draught Floating Units (LRFD)

1.6 Symbols

A Ae As Ap Aw B CD D E Ep Ekin Es F G H I J Kl Km K lm L M MP NP Sd T N NSd NRd NP R RD R0 V WP W a as ai b bf c cf Cross-sectional area Effective area of stiffener and effective plate flange Area of stiffener Projected cross-sectional area Shear area of stiffener/girder Width of contact area Hydrodynamic drag coefficient Diameter of circular sections, plate stiffness Young's Modulus of elasticity, (for steel 2.1105 N/mm2) Plastic modulus Kinetic energy Strain energy Lateral load, total load Shear modulus Non-dimensional plastic stiffness Moment of inertia, impulse Mass moment of inertia Load transformation factor Mass transformation factor Load-mass transformation factor Girder length Total mass, cross-sectional moment Plastic bending moment resistance Plastic axial resistance Design load effect Fundamental period of vibration Axial force Design axial compressive force Design axial compressive capacity Axial resistance of cross section Resistance Design resistance Plastic collapse resistance in bending Volume, displacement Plastic section modulus Elastic section modulus Added mass Added mass for ship Added mass for installation Width of collision contact zone Flange width Factor Axial flexibility factor

c lp cs cQ cw d dc fu fy g hw i k

k ke

Plastic zone length factor Shear factor for vibration eigenperiod Shear stiffness factor Displacement factor for strain calculation Smaller diameter of threaded end of drill collar Characteristic dimension for strain calculation Generalised load Ultimate material tensile strength Characteristic yield strength Acceleration of gravity, 9.81 m/s2 Web height for stiffener/girder Radius of gyration Stiffness, characteristic stiffness, plate stiffness, factor Generalised stiffness Equivalent stiffness Bending stiffness in linear domain for beam Stiffness in linear domain including shear deformation Shear stiffness in linear domain for beam Temperature reduction of effective yield stress for maximum temperature in connection Plate length, beam length Distributed mass Ship mass Installation mass Equivalent mass Generalised mass Explosion pressure Radius of deformed area, resistance Plastic collapse resistance in bending for plate Radius of gyration Distance, stiffener spacing Characteristic distance Effective width of plate Thickness, time Duration of explosion Flange thickness Web thickness Velocity of ship Velocity of installation Terminal velocity Deformation, displacement Characteristic deformation dent depth Non-dimensional deformation Axial coordinate

kl

k

' 1

kQ

k y,

l m

ms mi meq

m

p r rc rg s sc se t td tf tw vs vi vt w wc wd

w

x

y yel z

cr y

Generalised displacement, displacement amplitude Generalised displacement at elastic limit Distance from pivot point to collision point Plate aspect parameter Cross-sectional slenderness factor Yield strength factor, strain Critical strain for rupture Yield strain Plate eigenperiod parameter Displacement shape function Reduced slenderness ratio Ductility ratio Poisson's ratio, 0.3 Angle Density of steel, 7860 Shear stress Critical shear stress for plate plugging Interpolation factor Plate stiffness parameter kg/m3 Density of sea water, 1025 kg/m3

The requirements to structures exposed for accidental loads are given in DNV-OS-C101 Section 7. The structure should be checked in two steps: First the structure will be checked for the loads to which it is exposed due to the accidental event Secondly in case the structural capacity towards ordinary loads is reduced as a result of the accident then the strength of the structure is to be rechecked for ordinary loads. The structure should be checked for all relevant limit states. The limit states for accidental loads are denoted Accidental Limit States (ALS). The requirement may be written as

w cr

Sd R d

where: Sd Rd Sk f R M = = = = = =

(2.1)

Rk M

Design resistance Characteristic load effect partial factor for loads Characteristic resistance Material factor

For check of Accidental limit states (ALS) the load and material factor should be taken as 1.0.

2. Design Philosophy

2.1 General

The overall goal for the design of the structure against accidental loads is to prevent an incident to develop into an accident disproportional to the original cause. This means that the main safety functions should not be impaired by failure in the structure due to the design accidental loads. With the main safety functions is understood: usability of escapeways, integrity of shelter areas, global load bearing capacity In this section the design procedure that is intended to fulfil this goal is presented. The design against accidental loads may be done by direct calculation of the effects imposed on the structure, or indirectly, by design of the structure as tolerable to accidents. Examples of the latter are compartmentation of floating units which provides sufficient integrity to survive certain collision scenarios without further calculations. The inherent uncertainty of the frequency and magnitude of the accidental loads, as well as the approximate nature of the methods for determination of accidental load effects, shall be recognised. It is therefore essential to apply sound engineering judgement and pragmatic evaluations in the design. Typical accidental events are: Ship collision Dropped objects Fire Explosion

The failure criterion needs to be seen in conjunction with the assumptions made in the safety evaluations. The limit states may need to be alternatively formulated to be on the form of energy formulation, as acceptable deformation, or as usual on force or moment.

The accidental loads are either prescriptive values or defined in a Formal Safety Assessment. Prescriptive values may be given by authorities, the owner or found in DNV Offshore Standard DNV-OS-A101. Usually the simplification that accidental loads need not to be combined with environmental loads is valid. For check of the residual strength in cases where the accident lead to reduced load carrying capacity in the structure the check should be made with the characteristic environmental loads determined as the most probable annual maximum value.

Examples of failure criteria are: Critical deformation criteria defined by integrity of passive fire protection. To be considered for walls resisting explosion pressure and shall serve as fire barrier after the explosion. Critical deflection for structures to avoid damage to process equipment (Riser, gas pipe, etc). To be considered for structures or part of structures exposed to impact loads as ship collision, dropped object etc. Critical deformation to avoid leakage of compartments. To be considered in case of impact against floating structures where the acceptable collision damage is defined by the minimum number of undamaged compartments to remain stable.

The mechanical response to accidental loads is generally concerned with energy dissipation, involving large deformations and strains far beyond the elastic range. Hence, load effects (stresses forces, moments etc.) obtained from elastic analysis and used in ultimate limit state (ULS) checks on component level are generally not applicable, and plastic methods of analysis should be used. Plastic analysis is most conveniently based upon the kinematical approach, taking into account the effect of the strengthening (membrane tension) or softening (compression) caused by finite deformations, where applicable. The requirements in this RP are generally derived from plastic methods of analysis, including the effect of finite deformations. Plastic methods of analysis are valid for materials that can undergo considerable straining and during this process exhibit considerable strain hardening. If the material is ductile as such, i.e. it can be strained significantly, but has little strain hardening, the member tends to behave brittle in a global sense (i.e. with respect to energy dissipation), and plastic methods should be used with great caution. A further condition for application of plastic methods to members undergoing large, plastic rotations is compact cross-sections; typically type I cross-sections (refer DNV-OS-C101, Table A1). The methods may also be utilised for type II sections provided that the detrimental effect of local buckling is taken into account. Note that for members subjected to significant tensile straining, the tendency for local buckling may be overridden by membrane tension for large deformations. The straining, and hence the amount of energy dissipation, is limited by fracture. This key parameter is associated with considerable uncertainty, with respect to both physical occurrence as well as modelling in theoretical analysis. If good and validated models for prediction of fracture are not available, safe and conservative assumptions for ductility limits should be adopted. If non-linear, dynamic finite elements analysis is applied, it shall be verified that all behavioural effects and local failure modes (e.g. strain rate, local buckling, joint overloading, and joint fracture) are accounted for implicitly by the modelling adopted, or else subjected to explicit evaluation.

for the installation and the ship, where the deformations in the installation shall comply with ductility and stability requirements. The load bearing function of the installation shall remain intact with the damages imposed by the ship collision load. In addition, damaged condition should be checked if relevant, see Section 2.2. The structural effects from ship collision may either be determined by non-linear dynamic finite element analyses or by energy considerations combined with simple elastic-plastic methods. If non-linear dynamic finite element analysis is applied all effects described in the following paragraphs shall either be implicitly covered by the modelling adopted or subjected to special considerations, whenever relevant. Often the integrity of the installation can be verified by means of simple calculation models. If simple calculation models are used the part of the collision energy that needs to be dissipated as strain energy can be calculated by means of the principles of conservation of momentum and conservation of energy, refer Section 3.3. It is convenient to consider the strain energy dissipation in the installation to take part on three different levels: local cross-section component/sub-structure total system Interaction between the three levels of energy dissipation shall be considered. Plastic modes of energy dissipation shall be considered for cross-sections and component/substructures in direct contact with the ship. Elastic strain energy can in most cases be disregarded, but elastic axial flexibility may have a substantial effect on the load-deformation relationships for components/ sub-structures. Elastic energy may contribute significantly on a global level.

With respect to the distribution of strain energy dissipation there may be distinguished between, see Figure 3-1: strength design ductility design shared-energy design

Energy dissipation Ductile design Shared-energy design Strength design

ship

3. Ship Collisions

3.1 General

The requirements and methods given in this section have historically been developed for jackets. They are generally valid also for jack-up type platforms, provided that the increased importance of global inertia effects are accounted for. Columnstabilised platforms and floating production and storage vessels (FPSOs) consist typically plane or curved, stiffened panels, for which methods for assessment of energy dissipation in braced platforms (jackets and jack-ups) sometimes are not relevant. Procedures especially dedicated to assessment of energy dissipation in stiffened plating are, however, also given based on equivalent beam-column models. The ship collision load is characterised by a kinetic energy, governed by the mass of the ship, including hydrodynamic added mass and the speed of the ship at the instant of impact. Depending upon the impact conditions, a part of the kinetic energy may remain as kinetic energy after the impact. The remainder of the kinetic energy has to be dissipated as strain energy in the installation and, possibly, in the vessel. Generally this involves large plastic strains and significant structural damage to the installation, the ship or both. The strain energy dissipation is estimated from force-deformation relationships

Figure 3-1 Energy dissipation for strength, ductile and shared-energy design

Strength design implies that the installation is strong enough to resist the collision force with minor deformation, so that the ship is forced to deform and dissipate the major part of the energy. Ductility design implies that the installation undergoes large, plastic deformations and dissipates the major part of the collision energy.

Shared energy design implies that both the installation and ship contribute significantly to the energy dissipation. From calculation point of view strength design or ductility design is favourable. In this case the response of the soft structure can be calculated on the basis of simple considerations of the geometry of the rigid structure. In shared energy design both the magnitude and distribution of the collision force depends upon the deformation of both structures. This interaction makes the analysis more complex. In most cases ductility or shared energy design is used. However, strength design may in some cases be achievable with little increase in steel weight.

Collision response

Model

3.3.1 Strain energy dissipation The collision energy to be dissipated as strain energy may depending on the type of installation and the purpose of the analysis - be taken as: Compliant installations

vi 1 v 1 2 s E s = (ms + a s )vs ms + a s 2 1+ mi + a i

2

The structural response of the ship and installation can formally be represented as load-deformation relationships as illustrated in Figure 3-3. The strain energy dissipated by the ship and installation equals the total area under the load-deformation curves.

(3.1)

Rs

Ri

Fixed installations

1 2 E s = (m s + a s )v s 2

(3.2)

Es,i

Articulated columns

v 1 i vs 1 E s = (m s + a s ) 2 ms z 2 1+ J

2

Installation

dwi

(3.3)

ms as vs mi ai vi J z

= = = = = = =

ship mass ship added mass impact speed mass of installation added mass of installation velocity of installation mass moment of inertia of installation (including added mass) with respect to effective pivot point = distance from pivot point to point of contact

E s = E s,s + E s,i =

w s, max 0

R s dw s +

w i, max 0

R i dw i

(3.4)

In most cases the velocity of the installation can be disregarded, i.e. vi = 0. The installation can be assumed compliant if the duration of impact is small compared to the fundamental period of vibration of the installation. If the duration of impact is comparatively long, the installation can be assumed fixed. Floating platforms (semi-submersibles, TLPs, production vessels) can normally be considered as compliant. Jack-ups may be classified as fixed or compliant. Jacket structures can normally be considered as fixed.

3.3.2 Reaction force to deck In the acceleration phase the inertia of the topside structure generates large reaction forces. An upper bound of the maximum force between the collision zone and the deck for bottom supported installations may be obtained by considering the platform compliant for the assessment of total strain energy dissipation and assume the platform fixed at deck level when the collision response is evaluated.

As the load level is not known a priori an incremental procedure is generally needed. The load-deformation relationships for the ship and the installation are often established independently of each other assuming the other object infinitely rigid. This method may have, however, severe limitations; both structures will dissipate some energy regardless of the relative strength. Often the stronger of the ship and platform will experience less damage and the softer more damage than what is predicted with the approach described above. As the softer structure deforms the impact force is distributed over a larger contact area. Accordingly, the resistance of the strong structure increases. This may be interpreted as an "upward" shift of the resistance curve for the stronger structure (refer Figure 3-3 ). Care should be exercised that the load-deformation curves calculated are representative for the true, interactive nature of the contact between the two structures.

3.5.1 Recommended force-deformation relationships Force-deformation relationships for supply vessels with a displacement of 5000 tons are given in Figure 3-4 for broad side -, bow-, stern end and stern corner impact for a vessel with stern roller. The curves for broad side and stern end impacts are based upon

penetration of an infinitely rigid, vertical cylinder with a given diameter and may be used for impacts against jacket legs (D = 1.5 m) and large diameter columns (D = 10 m). The curve for stern corner impact is based upon penetration of an infinitely rigid cylinder and may be used for large diameter column impacts. In lieu of more accurate calculations the curves in Figure 3-4 may be used for square-rounded columns. The curve for bow impact is based upon collision with an infinitely rigid, plane wall and may be used for large diameter column impacts, but should not be used for significantly different

collision events, e.g. impact against tubular braces. For beam -, stern end and stern corner impacts against jacket braces all energy shall normally be assumed dissipated by the brace, refer Ch.8, Comm. 3.5.2. For bow impacts against jacket braces, reference is made to Section 3.5.3. For supply vessels and merchant vessels in the range of 25000 tons displacement, the force deformation relationships given in Figure 3-5 may be used for impacts against jacket legs with diameter 1.5 m 2.5 m.

50 40 Impact force (MN) 30 20 10 0 0 1 Bo Bow 2 Indentation (m) 3 4 Stern corner Stern end D = 10 m = 1.5 m

D

Figure 3-4 Recommended-deformation curve for beam, bow and stern impact

Figure 3-5 Force -deformation relationship for bow with and without bulb (2-5.000 dwt)

40 Design curve - plane wall Energy Contact force with bulb 30 20 10 0 Force [MN]

70 60

12 10

Force [MN]

50 40 30 20 10 0 0 1 2 3 4

b a

b a

5 6

8 6 4 2 0

Bulb force

Force-deformation relationships for tanker bow impact are given in Figure 3-6 for the bulbous part and the superstructure, respectively, and for the bulb of a VLCC in Figure 3-7. The curves may be used provided that the impacted structure (e.g. stern of floating production vessels) does not undergo substantial deformation i.e. strength design requirements are complied with. If this condition is not met interaction between the bow and the impacted structure shall be taken into consideration. Non-linear finite element methods or simplified plastic analysis techniques of members subjected to axial crushing shall be employed, see Ch.7 /3/, /4/.

3.5.2 Force contact area for strength design of large diameter columns.

Deformation [m]

70 60

18

Force [MN]

50 40 30 20 10 0 0 1

14

a b

b Force superstructure

2 3

12 10 8 6 4 2 0

16

The basis for the curves in Figure 3-4 is strength design, i.e. limited local deformations of the installation at the point of contact. In addition to resisting the total collision force, large diameter columns have to resist local concentrations (subsets) of the collision force, given for stern corner impact in Table 31 and stern end impact in Table 3-2.

Table 3-1 Local concentrated collision force -evenly distributed over a rectangular area. Stern corner impact Contact area Force (MN) a (m) b (m) b 0.35 0.65 3.0 0.35 1.65 6.4 a 0.20 1.15 5.4 Table 3-2 Local concentrated collision force -evenly distributed over a rectangular area. Stern end impact Contact area Force (MN) a (m) b (m) b 0.6 0.3 5.6 a 0.9 0.5 7.5 2.0 1.1 10

a

4 5

Deformation [m]

Figure 3-6 Force -deformation relationship for tanker bow impact (~ 125.000 dwt)

160 140 120 Force [MN] 100 80 60 40 20 0 0 1 2 3 4 5 6 Deformation [m] 7 8 Force Energy

800 700 500 400 300 200 100 0 Energy [MJ] 600

If strength design is not aimed for - and in lieu of more accurate assessment (e.g. nonlinear finite element analysis) - all strain energy has to be assumed dissipated by the column, corresponding to indentation by an infinitely rigid stern corner.

3.5.3 Energy dissipation is ship bow

a b

For typical supply vessels bows and bows of merchant vessels of similar size (i.e. 2-5000 tons displacement), energy dissipation in ship bow may be taken into account provided that the collapse resistance in bending for the brace, R0, see Section 3.7 is according to the values given in Table 3-3. The figures are valid for normal bows without ice strengthening and for brace diameters < 1.25 m. The values should be used as step functions, i.e. interpolation for intermediate resistance levels is not allowed. If contact location is not governed by operation conditions, size of ship and platform etc., the values for arbitrary contact location shall be used. (see also Ch.8, Comm. 3.5.3).

Table 3-3 Energy dissipation in bow versus brace resistance Energy dissipation in bow Contact location if brace resistance R0 > 3 MN > 6 MN > 8 MN > 10 MN Above bulb 1 MJ 4 MJ 7 MJ 11 MJ First deck 0 MJ 2 MJ 4 MJ 17 MJ First deck - oblique brace 0 MJ 2 MJ 4 MJ 17 MJ Between forcastle/first 1 MJ 5 MJ 10 MJ 15 MJ deck Arbitrary location 0 MJ 2 MJ 4 MJ 11 MJ

b a

Figure 3-7 Force -deformation relationship and contact area for the bulbous bow of a VLCC (~ 340.000 dwt)

compactness requirement

f y t 1.5 D 0.5 2 factor 3

(3.5)

where factor is the required resistance in [MN] given in Table 3-3. See Section 3.6 for notation. If the brace is designed to comply with these provisions, special care should be exercised that the joints and adjacent structure is strong enough to support the reactions from the brace.

The curves are inaccurate for small indentation, and they should not be used to verify a design where the dent damage is required to be less than wd / D > 0.05. The width of contact area is in theory equal to the height of the vertical, plane section of the ship side that is assumed to be in contact with the tubular member. For large widths, and depending on the relative rigidity of the cross-section and the ship side, it may be unrealistic to assume that the tube is subjected to flattening over the entire contact area. In lieu of more accurate calculations it is proposed that the width of contact area be taken equal to the diameter of the hit cross-section (i.e. B/D = 1).

The contribution from local denting to energy dissipation is small for brace members in typical jackets and should be neglected. The resistance to indentation of unstiffened tubes may be taken from Figure 3-8. Alternatively, the resistance may be calculated from Equation (3.6):

20 18 16 14

3.7.1 General The response of a beam subjected to a collision load is initially governed by bending, which is affected by and interacts with local denting under the load. The bending capacity is also reduced if local buckling takes place on the compression side. As the beam undergoes finite deformations, the load carrying capacity may increase considerably due to the development of membrane tension forces. This depends upon the ability of adjacent structure to restrain the connections at the member ends to inward displacements. Provided that the connections do not fail, the energy dissipation capacity is either limited by tension failure of the member or rupture of the connection. Simple plastic methods of analysis are generally applicable. Special considerations shall be given to the effect of:

b/D =

R/(kRc)

12 10 8 6 4 2 0 0 0.1

2 1 0.5 0

0.2

0.3

0.4

0.5

elastic flexibility of member/adjacent structure, local deformation of cross-section, local buckling, strength of connections, strength of adjacent structure, and fracture.

wd/D

Figure 3-8 Resistance curve for local denting

3.7.2 Plastic force-deformation relationships including elastic, axial flexibility Relatively small axial displacements have a significant influence on the development of tensile forces in members undergoing large lateral deformations. An equivalent elastic, axial stiffness may be defined as

R wd = kc1 D Rc Rc = fy t2 4

c2

D t B c1 = 22 + 1.2 D 1.925 c2 = B 3.5 + D k = 1.0 N Sd k = 1.0 2 N 0.2 Rd k=0 N Sd 0.2 N Rd 0.2 < N Sd < 0.6 N Rd 0.6 N Sd N Rd

l 1 1 = + k k node 2EA

(3.7)

(3.6)

knode = axial stiffness of the node with the considered member removed. This may be determined by introducing unit loads in member axis direction at the end nodes with the member removed. Plastic force-deformation relationship for a central collision (midway between nodes) may be obtained from: Figure 3-9 for tubular members Figure 3-10 for stiffened plates in lieu of more accurate analysis. The following notation applies:

R0 = 4c1M P l

plastic collapse resistance in bending for the member, for the case that contact point is at midspan non-dimensional deformation

w=

NSd NRd B wd

= = = =

design axial compressive force design axial compressive resistance width of contactarea dent depth

w c1 wc

2

c=

4c 1 kw c f y Al

c1 = 2

c1 = 1

wc =

wc =

WP

D 2

1 .2 WP A

= =

characteristic deformation for tubular beams characteristic deformation for stiffened plating plastic section modulus member length

member the force-deformation relationship may be interpolated from the curves for pinned ends and clamped ends:

R = R clamped + (1 ) R pinned

where

(3.8)

0=

For non-central collisions the force-deformation relationship may be taken as the mean value of the force-deformation curves for central collision with member half length equal to the smaller and the larger portion of the member length, respectively. For members where the plastic moment capacity of adjacent members is smaller than the moment capacity of the impacted member the force-deformation relationship may be interpolated from the curves for pinned ends and clamped ends: For non-central collisions the force-deformation relationship may be taken as the mean value of the force-deformation curves for central collision with member half length equal to the smaller and the larger portion of the member length, respectively. For members where the plastic moment capacity of adjacent members is smaller than the moment capacity of the impacted

6,5 6 5,5 5 4,5 4

R0 1 1 MP 4 l

actual

(3.9)

R0

R0

actual

resistance by bending action of beam account= Plastic ing for actual bending resistance of adjacent members

actual

=

i

M Pj = M Pj,i M P

4M P + 2M P1 + 2M P2 l

(3.10) (3.11)

i = adjacent member no i j = end number {1,2} MPj,i = Plastic bending resistance for member number i at end j. Elastic, rotational flexibility of the node is normally of moderate significance.

k

R/R0

(collision load)

c =

1 0.05

2,5

3,5

Deformation

Figure 3-9 Force-deformation relationship for tubular beam with axial flexibility

5 4,5 4 3,5 3

(collision load)

R/R0

c =

0.5 1

0.2

w

0

3,5

Deformation

Figure 3-10 Force-deformation relationship for stiffened plate with axial flexibility

3.7.3 Support capacity smaller than plastic bending moment of the beam

For beams where the plastic moment capacity of adjacent members is smaller than the moment capacity of the impacted beam, the force-deformation relationship, R*, may be derived from the resistance curve, R, for beams where the plastic moment capacity of adjacent members is larger than the moment capacity of the impacted beam (Section 3.7.2), using the expression:

i = j = MPj,i= wlim =

adjacent member no i end number {1,2} Plastic bending resistance for member no. i limiting non-dimensional deformation where the membrane force attains yield, i.e. the resistance curve, R, with actual spring stiffness coefficient, c, intersects with the curve for c = . If c = , w lim = w for 2 tubular beams and wlim = 1.2 w for stiffened plate

R = R + (R 0 R )

* * 0

w w lim

w w lim

1 .0

(3.12)

R* = R

w w lim

1.0

where R0 = Plastic bending resistance with clamped ends (c1 = 2) moment capacity of adjacent members larger than the plastic bending moment of the beam * R 0 = Plastic bending resistance - moment capacity of adjacent members at one or both ends smaller than the plastic bending moment of the beam

4M P + 2M P1 + 2M P2 R* 0 = l

M Pj = M Pj,i M P

i

3.7.4 Bending capacity of dented tubular members The reduction in plastic moment capacity due to local denting shall be considered for members in compression or moderate tension, but can be neglected for members entering the fully plastic membrane state. Conservatively, the flat part of the dented section according to the model shown in Figure 3-11 may be assumed non-effective. This gives:

M red 1 = cos sin MP 2 2 M P = f y D2 t 2w d = arccos1 D

(3.15)

(3.13)

(3.14)

wd D

occur when the tensile strain due to the combined effect of rotation and membrane elongation exceeds a critical value. To ensure that members with small axial restraint maintain moment capacity during significant plastic rotation it is recommended that cross-sections be proportioned to section type I requirements, defined in DNV-OS-C101. Initiation of local buckling does, however, not necessarily imply that the capacity with respect to energy dissipation is exhausted, particularly for type I and type II cross-sections. The degradation of the cross-sectional resistance in the postbuckling range may be taken into account provided that such information is available, refer Ch.8, Comm. 3.10.1. For members undergoing membrane stretching a lower bound to the post-buckling load-carrying capacity may be obtained by using the load-deformation curve for pure membrane action.

3.10.2 Local buckling Tubular cross-sections: Buckling does not need to be considered for a beam with axial restraints if the following condition is fulfilled:

14c f f y c1 l d c

2

Mred/MP

wd/D

Figure 3-11 Reduction of moment capacity due to local dent

Provided that large plastic strains can develop in the impacted member, the strength of the connections that the member frames into should be checked. The resistance of connections should be taken from ULS requirements in relevant standards. For braces reaching the fully plastic tension state, the connection shall be checked for a load equal to the axial capacity of the member. The design axial stress shall be assumed equal to the ultimate tensile strength of the material. If the axial force in a tension member becomes equal to the axial capacity of the connection, the connection has to undergo gross deformations. The energy dissipation will be limited and rupture should be considered at a given deformation. A safe approach is to assume failure (disconnection of the member) once the axial force in the member reaches the axial capacity of the connection. If the capacity of the connection is exceeded in compression and bending, this does not necessarily mean failure of the member. The post-collapse strength of the connection may be taken into account provided that such information is available.

(3.16)

where

= Dt 235 f y

(3.17)

c cf = 1 + c

2

(3.18)

dc c1 c

= = = = =

characteristic dimension D for circular cross-sections 2 for clamped ends 1 for pinned ends non-dimensional spring stiffness, refer Section 3.7.2.

The strength of structural members adjacent to the impacted member/sub-structure must be checked to see whether they can provide the support required by the assumed collapse mechanism. If the adjacent structure fails, the collapse mechanism must be modified accordingly. Since, the physical behaviour becomes more complex with mechanisms consisting of an increasing number of members it is recommended to consider a design which involves as few members as possible for each collision scenario.

l 0.5 l = the smaller distance from location of collision load to adjacent joint If this condition is not met, buckling may be assumed to occur when the lateral deformation exceeds

14c f f y w 1 = 1 1 d c 2c f c 1 3

2 l d c

(3.19)

For small axial restraint (c < 0.05) the critical deformation may be taken as

w 3.5f y = dc c 1 3 l d c

2

3.10.1 General The maximum energy that the impacted member can dissipate will ultimately - be limited by local buckling on the compressive side or fracture on the tensile side of cross-sections undergoing finite rotation. If the member is restrained against inward axial displacement, any local buckling must take place before the tensile strain due to membrane elongation overrides the effect of rotation induced compressive strain. If local buckling does not take place, fracture is assumed to

(3.20)

Stiffened plates/ I/H-profiles: In lieu of more accurate calculations the expressions given for circular profiles in Equation (3.19) and (3.20) may be used with dc = characteristic dimension for local buckling, equal to twice the distance from the plastic neutral axis in bending to the extreme fibre of the cross-section = h height of cross-section for symmetric I profiles = 2hw for stiffened plating (for simplicity)

= 2.5

=3

bf t f 235 f y

235 f y

(3.21) (3.22)

3.10.4 Tensile fracture in yield hinges When the force deformation relationships for beams given in Section 3.7.2 are used rupture may be assumed to occur when the deformation exceeds a value given by

bf t f

= 0.7

= 0.8

4c c c w = 1 1 + w f cr 1 d c 2c f c1

where the following factors are defined; Displacement factor

(3.26)

hw tw 235 f y

hw tw 235 f y

(3.23) (3.24)

1 cw = c1

1 c lp 1 c lp + 41 W 3 W P

y cr

l d c

(3.27)

bf tf hw tw

= = = =

cr W 1 H y WP c lp = cr W 1 H +1 y WP

3.10.3 Tensile fracture The degree of plastic deformation or critical strain at fracture will show a significant scatter and depends upon the following factors:

(3.28)

c cf = 1+ c

2

(3.29)

The critical strain for plastic deformations of sections containing defects need to be determined based on fracture mechanics methods. Welds normally contain defects and welded joints are likely to achieve lower toughness than the parent material. For these reasons structures that need to undergo large plastic deformations should be designed in such a way that the plastic straining takes place outside the weld. In ordinary full penetration welds, the overmatching weld material will ensure that minimal plastic straining occurs in the welded joints even in cases with yielding of the gross cross section of the member. In such situations, the critical strain will be in the parent material and will be dependent upon the following parameters: stress gradients dimensions of the cross section presence of strain concentrations material yield to tensile strength ratio material ductility

H= Ep E = f fy 1 cr E cr y

(3.30)

c1 c l W WP cr

= = = = = =

2 for clamped ends 1 for pinned ends non-dimensional spring stiffness, refer Section 3.7.2 0.5 l the smaller distance from location of collision load to adjacent joint elastic section modulus plastic section modulus critical strain for rupture (see Table 3-4 for recommended values) yield strain yield strength strength corresponding to cr

Simple plastic theory does not provide information on strains as such. Therefore, strain levels should be assessed by means of adequate analytic models of the strain distributions in the plastic zones or by non-linear finite element analysis with a sufficiently detailed mesh in the plastic zones. (For information about mesh size see Ch.8, Comm. 3.10.4.) When structures are designed so that yielding take place in the parent material, the following value for the critical average strain in axially loaded plate material may be used in conjunction with nonlinear finite element analysis or simple plastic analysis

y =

fy fcr

fy E

= = =

The characteristic dimension shall be taken as: dc = D = 2hw = h = 2 (h zplast) diameter of tubular beams twice the web height for stiffened plates (set > As) height of cross-section for symmetric I-profiles for unsymmetrical I-profiles

cr = 0.02 + 0.65

where: t

t l

(3.25)

zplast = smaller distance from flange to plastic neutral axis of cross-section For small axial restraint (c < 0.05) the critical deformation may

be taken as

w = c w cr dc

(3.31)

effective shell flange WP = plastic section modulus of ring stiffener including effective shell flange Ae = area of ring stiffener including effective shell flange Effective flange of shell plating: Use effective flange of stiffened plates, see Chapter 6. For ductile design it can be assumed that the resistance of the ring stiffener is constant and equal to the plastic collapse load, provided that requirements for stability of cross-sections are complied with, refer Section 3.10.2.

3.11.4 Decks and bulkheads Calculation of energy dissipation in decks and bulkheads has to be based upon recognised methods for plastic analysis of deep, axial crushing. It shall be documented that the collapse mechanisms assumed yield a realistic representation of the true deformation field.

The critical strain cr and corresponding strength fcr should be selected so that idealised bi-linear stress-strain relation gives reasonable results, see Ch.8, Commentary. For typical steel material grades the following values are proposed:

Table 3-4 Proposed values for cr and H for different steel grades Steel grade cr H S 235 20 % 0.0022 S 355 15 % 0.0034 S 460 10 % 0.0034

3.11.1 General Impact on a ring stiffener as well as midway between ring stiffeners shall be considered. Plastic methods of analysis are generally applicable. 3.11.2 Longitudinal stiffeners For ductile design the resistance of longitudinal stiffeners in the beam mode of deformation can be calculated using the procedure described for stiffened plating, Section 3.7. For strength design against stern corner impact, the plastic bending moment capacity of the longitudinal stiffeners has to comply with the requirement given in Figure 3-12, on the assumption that the entire load given in Table 3-1 is taken by one stiffener.

3

For strength design the side or stern shall resist crushing force of the bow of the off-take tanker. In lieu of more accurate calculations the force-deformation curve given in Section 3.5.2 may be applied. (See Ch.8, Comm. 3.12 on strength design of stern structure) For ductile design the resistance of stiffened plating in the beam mode of deformation can be calculated using the procedure described in Section 3.7.2. (See Ch.8, Comm. 3.12 on resistance of stiffened plating)

Normally, it is unlikely that the installation will turn into a global collapse mechanism under direct collision load, because the collision load is typically an order of magnitude smaller than the resultant design wave force. Linear analysis often suffices to check that global integrity is maintained. The installation should be checked for the maximum collision force. For installations responding predominantly statically the maximum collision force occurs at maximum deformation. For structures responding predominantly impulsively the maximum collision force occurs at small global deformation of the platform. An upper bound to the collision force is to assume that the installation is fixed with respect to global displacement. (e.g. jack-up fixed with respect to deck displacement).

0 1 2 3 4

4. Dropped Objects

4.1 General

The dropped object load is characterised by a kinetic energy, governed by the mass of the object, including any hydrodynamic added mass, and the velocity of the object at the instantof impact. In most cases the major part of the kinetic energy has to be dissipated as strain energy in the impacted component and, possibly, in the dropped object. Generally, this involves large plastic strains and significant structural damage to the impacted component. The strain energy dissipation is estimated from force-deformation relationships for the component and the object, where the deformations in the component shall comply with ductility and stability requirements. The load bearing function of the installation shall remain intact with the damages imposed by the dropped object load. In addition, damaged condition should be checked if relevant, see Section 2.2. Dropped objects are rarely critical to the global integrity of the

3.11.3 Ring stiffeners In lieu of more accurate analysis the plastic collapse load of a ring-stiffener can be estimated from:

F0 =

4 2M P

wc D

(3.32)

where

wc = WP Ae

installation and will mostly cause local damages. The major threat to global integrity is probably puncturing of buoyancy tanks, which could impair the hydrostatic stability of floating installations. Puncturing of a single tank is normally covered by the general requirements to compartmentation and watertight integrity given in DNV-OS-C301. The structural effects from dropped objects may either be determined by non-linear dynamic finite element analyses or by energy considerations combined with simple elastic-plastic methods as given in Sections 4.2 - 4.5. If non-linear dynamic finite element analysis is applied all effects described in the following paragraphs shall either be implicitly covered by the modelling adopted or subjected to special considerations, whenever relevant.

and

E kin = 1 (m + a )v 2 (in water) 2

(4.2)

v = 2gs

(4.3)

s v

The kinetic energy of a falling object is given by:

E kin = 1 mv 2 2

(in air)

(4.1)

For objects falling rectilinearly in water the velocity depends upon the reduction of speed during impact with water and the falling distance relative to the characteristic distance for the object.

3 3,5 4

Distance [s/sc]

1 2 3 4 5 6 7

In water

t

(4.4)

F(t) = force during impact with sea surface After the impact with water the object proceeds with the speed v = v 0 v Assuming that the hydrodynamic resistance during fall in water is of drag type the velocity in water can be taken from Figure 4-1 where

m+a sc = = w Cd A p

w = density of sea water Cd = hydrodynamic drag coefficient for the object in the considered motion m = mass of object Ap = projected cross-sectional area of the object V = object displacement The major uncertainty is associated with calculating the loss of momentum during impact with sea surface, given by Equation (4.4). However, if the travelled distance is such that the velocity is close to the terminal velocity, the impact with sea surface is of little significance. Typical terminal velocities for some typical objects are given

2g(m w V) vt = w Cd A p

terminal velocity for the = object (drag force and buoyancy force balance the gravity force)

in Table 4-1.

Table 4-1 Terminal velocities for objects falling in water Item Mass Terminal velocity [kN] [m/s] Drill collar 28 23-24 Winch, 250 Riser pump 100 BOP annular preventer 50 16 Mud pump 330 7

c = e

d 2.5 1 2r

m i = p r 2 t

= mass of plate enclosed by hinge circle = mass of dropped object = mass density of steel plate = smaller diameter at threaded end of drill collar = smaller distance from the point of impact to the plate boundary defined by adjacent stiffeners/girders, refer Figure 4-3.

Rectilinear motion is likely for blunt objects and objects which do not rotate about their longitudinal axis. Bar-like objects (e.g. pipes) which do not rotate about their longitudinal axis may execute lateral, damped oscillatory motions as illustrated in Figure 4-1.

m p d r

The structural response of the dropped object and the impacted component can formally be represented as load-deformation relationships as illustrated in Figure 4-2. The part of the impact energy dissipated as strain energy equals the total area under the load-deformation curves. E s = E s,o + E s,i =

w o, max 0

For validity range of design formula reference is given to Ch.8, Comm. 4.4.1.

R o dw o +

w i , max 0

R i dw i

(4.5)

As the load level is not known a priori an incremental approach is generally required. Often the object can be assumed to be infinitely rigid (e.g. axial impact from pipes and massive objects) so that all energy is to be dissipated by the impacted component.

Ro

Ri

4.4.2 Stiffeners/girders In lieu of more accurate calculations stiffeners and girders subjected to impact with blunt objects may be analysed with resistance models given in Section 6.10. 4.4.3 Dropped object Calculation of energy dissipation in deformable dropped objects shall be based upon recognised methods for plastic analysis. It shall be documented that the collapse mechanisms assumed yield a realistic representation of the true deformation field.

Es,i

Installation

dwi

4.5.1 Pipes on plated structures The maximum shear stress for plugging of plates due to drill collar impacts may be taken as

If the object is assumed to be deformable, the interactive nature of the deformation of the two structures should be recognised.

t cr = f u 0.42 + 0.41 d

f u = ultimate material tensile strength

(4.7)

4.4.1 Stiffened plates subjected to drill collar impact The energy dissipated in the plating subjected to drill collar impact is given by

4.5.2 Blunt objects For stability of cross-sections and tensile fracture, refer Section 3.10.

E sp

m R2 = 1 + 0.48 i 2k m

(4.6)

5. Fire

5.1 General

The characteristic fire structural load is temperature rise in exposed members. The temporal and spatial variation of temperature depends on the fire intensity, whether or not the structural members are fully or partly engulfed by the flame and to what extent the members are insulated. Structural steel expands at elevated temperatures and internal stresses are developed in redundant structures. These stresses

where:

d 1 + 5 d 6c 2 + 6.25 1 r 2r k = f y t 2 2 (1 + c)

are most often of moderate significance with respect to global integrity. The heating causes also progressive loss of strength and stiffness and is, in redundant structures, accompanied by redistribution of forces from members with low strength to members that retain their load bearing capacity. A substantial loss of load-bearing capacity of individual members and subassemblies may take place, but the load bearing function of the installation shall remain intact during exposure to the fire load. In addition, damaged condition should be checked if relevant, see Section 2.2. Structural analysis may be performed on either individual members subassemblies entire system The assessment of fire load effect and mechanical response shall be based on either simple calculation methods applied to individual members, general calculation methods, or a combination. Simple calculation methods may give overly conservative results. General calculation methods are methods in which engineering principles are applied in a realistic manner to specific applications. Assessment of individual members by means of simple calculation methods should be based upon the provisions given in Ch.7 /2/ Eurocode 3 Part 1.2. /2/ . Assessment by means of general calculation methods shall satisfy the provisions given in Ch.7 /2/ Eurocode 3 Part1.2, Section 4.3. In addition, the requirements given in this section for mechanical response analysis with nonlinear finite element methods shall be complied with. Assessment of ultimate strength is not needed if the maximum steel temperature is below 400C, but deformation criteria may have to be checked for impairment of main safety functions.

tions compressive members shall be modelled with an initial, sinusoidal imperfection with amplitude given by Elastic-perfectly plastic material model, refer Figure 6-4 :

e* 1 = l fy i E z0

e* W 1 = W l

p

fy i = E z0

f y Wp E AI

= 0.5 for fire exposed members according to column curve c, Ch.7 /2/ Eurocode 3 i = radius of gyration z0 = distance from neutral axis to extreme fibre of crosssection WP = plastic section modulus W = elastic section modulus A = cross-sectional area I = moment of inertia e* = amplitude of initial distortion l = member length The initial out-of-straightness should be applied on each physical member. If the member is modelled by several finite elements the initial out-of-straightness should be applied as displaced nodes. The initial out-of-straightness shall be applied in the same direction as the deformations caused by the temperature gradients.

The empirical correction factor of 1.2 should be accounted for in calculating the critical strength in compression and bending for design according to Ch.7 /2/ Eurocode 3, refer Ch.8, Comm. A.5.5.

If shell modelling is used, it shall be verified that the software and the modelling is capable of predicting local buckling with sufficient accuracy. If necessary, local shell imperfections have to be introduced in a similar manner to the approach adopted for lateral distortion of beams If beam modelling is used local cross-sectional buckling shall be given explicit consideration. In lieu of more accurate analysis cross-sections subjected to plastic deformations shall satisfy compactness requirements given in DNV-OS-C101: type I: Locations with plastic hinges (approximately full plastic utilization) type II: Locations with yield hinges (partial plastification) If this criterion is not complied with explicit considerations shall be performed. The load-bearing capacity will be reduced significantly after the onset of buckling, but may still be significant. A conservative approach is to remove the member from further analysis. Compactness requirements for type I and type I cross-sections may be disregarded provided that the member is capable of developing significant membrane forces.

Structural analysis methods for non-linear, ultimate strength assessment may be classified as stress-strain based methods stress-resultants based (yield/plastic hinge) methods Stress-strain based methods are methods where non-linear material behaviour is accounted for on fibre level. Stress-resultants based methods are methods where non-linear material behaviour is accounted for on stress-resultants level based upon closed form solutions/interaction equation for cross-sectional forces and moments.

In stress-strain based analysis temperature dependent stressstrain relationships given in Ch.7 /2/ Eurocode 3, Part 1.2, Section 3.2 may be used. For stress resultants based design the temperature reduction of the elastic modulus may be taken as kE, according to Ch.7 / 2/ Eurocode 3. The yield stress may be taken equal to the effective yield stress, fy,. The temperature reduction of the effective yield stress may be taken as ky,. Provided that the above requirements are complied with creep does need explicit consideration.

5.7.1 General The ductility of beams and connections increase at elevated temperatures compared to normal conditions. Little informa-

To account for the effect of residual stresses and lateral distor-

tion exists.

5.7.2 Beams in bending In lieu of more accurate analysis requirements given in Section 3.10 shall be complied with. 5.7.3 Beams in tension In lieu of more accurate analysis an average elongation of 3% of the member length with a reasonably uniform temperature can be assumed. Local temperature peaks may localise plastic strains. It is considered to be to the conservative side to use critical strains for steel under normal temperatures. See Section 3.10 and 3.10.4.

The load bearing function of the installation shall remain intact with the damages imposed by the explosion loads. In addition, damaged condition should be checked if relevant, see Section 2.2.

The response of structural components can conveniently be classified into three categories according to the duration of the explosion pressure pulse, td, relative to the fundamental period of vibration of the component, T: Impulsive domain Dynamic domain Quasi-static domain td/T < 0.3 0.3 < td/T < 3 3 < td/T

In lieu of more accurate calculations the capacity of the connection can be taken as: R = ky, R0 where R0 = capacity of connection at normal temperature ky, = temperature reduction of effective yield stress for maximum temperature in connection

I = F(t )dt

td 0

(6.1)

Hence, the structure may resist a very high peak pressure provided that the duration is sufficiently small. The maximum deformation, wmax, of the component can be calculated iteratively from the equation

I = 2m eq

w max 0

6. Explosions

6.1 General

Explosion loads are characterised by temporal and spatial pressure distribution. The most important temporal parameters are rise time, maximum pressure and pulse duration. For components and sub-structures the explosion pressure shall normally be considered uniformly distributed. On global level the spatial distribution is normally non-uniform both with respect to pressure and duration. The response to explosion loads may either be determined by non-linear dynamic finite element analysis or by simple calculation models based on Single Degree Of Freedom (SDOF) analogies and elastic-plastic methods of analysis. If non-linear dynamic finite element analysis is applied all effects described in the following paragraphs shall either be implicitly covered by the modelling adopted or subjected to special considerations, whenever relevant. In the simple calculation models the component is transformed to a single spring-mass system exposed to an equivalent load pulse by means of suitable shape functions for the displacements in the elastic and elastic-plastic range. The shape functions allow calculation of the characteristic resistance curve and equivalent mass in the elastic and elastic-plastic range as well as the fundamental period of vibration for the SDOF system in the elastic range. Provided that the temporal variation of the pressure can be assumed to be triangular, the maximum displacement of the component can be calculated from design charts for the SDOF system as a function of pressure duration versus fundamental period of vibration and equivalent load amplitude versus maximum resistance in the elastic range. The maximum displacement must comply with ductility and stability requirements for the component.

R (w )dw

(6.2)

where R(w)= force-deformation relationship for the component meq = equivalent mass for the component.

Quasi-static-domain: The response is governed by the peak pressure and the rise time of the pressure relative to the fundamental period of vibration. If the rise time is small the maximum deformation of the component can be solved iteratively from the equation:

w max = 1 Fmax

w max

R (w )dw

(6.3)

If the rise time is large the maximum deformation can be solved from the static condition

Fmax = R(w max )

(6.4)

Dynamic domain: The response has to be solved from numerical integration of the dynamic equations of equilibrium.

Various failure modes for a stiffened panel are illustrated in Figure 6-1. Suggested analysis model and reference to applicable resistance functions are listed in Table 6.1. Application of a Single Degree of Freedom (SDOF) model in the analysis of stiffeners/girders with effective flange is implicitly based on the assumption that dynamic interaction between the plate flange and the profile can be neglected.

Figure 6-1 Failure modes for two-way stiffened panel Table 6-1 Analysis models Failure mode Elastic-plastic deformation of plate Stiffener plastic plate elastic Stiffener plastic plate plastic Girder plastic stiffener and plating elastic Girder plastic stiffener elastic plate plastic Girder and stiffener plastic plate elastic Simplified analysis model Resistance models SDOF Section 6.9 SDOF Stiffener: Section 6.10.1-2. Plate: Section 6.9.1 SDOF Stiffener: Section 6.10.1-2. Plate: Section 6.9 SDOF Girder: Section 6.10.1-2 Plate: Section 6.9 SDOF Girder: Section 6.10.1-2 Plate: Section 6.9 Comment

Elastic, effective flange of plate Effective width of plate at mid span. Elastic, effective flange of plate at ends. Elastic, effective flange of plate with concentrated loads (stiffener reactions). Stiffener mass included. Effective width of plate at girder mid span and ends. Stiffener mass included

Girder and stiffener: Dynamic reactions of stiffeners loading for girders Section 6.10.1-2 Plate: Section 6.9 Girder and stiffener plastic MDOF Girder and stiffener: Dynamic reactions of stiffeners loading for girders plate plastic Section 6.10.1-2 Plate: Section 6.9 By girder/stiffener plastic is understood that the maximum displacement wmax exceeds the elastic limit wel

MDOF

Biggs method: For many practical design problems it is convenient to assume that the structure - exposed to the dynamic pressure pulse - ultimately attains a deformed configuration comparable to the static deformation pattern. Using the static deformation pattern as displacement shape function, i.e.

Kl = Kl =

q (x)dx

l

F F i

i i

load transformation factor for = uniformly distributed load = load transformation factor for concentrated load = total uniformly, distributed mass

F

l

w (x , t ) = (x )y(t )

the dynamic equations of equilibrium can be transformed to an equivalent single degree of freedom system:

M u = mdx

M

i

& + ky = f (t ) m& y

(x) y(t) m = m(x ) dx + M i i

2

l

(6.5) = displacement shape function = displacement amplitude = generalized mass = generalized load = generalized elastic bending stiffness = generalized plastic bending stiffness (fully developed mechanism) = generalized membrane stiffness (fully plastic: N = NP) = distributed mass = concentrated mass = explosion load = concentrated load (e.g. support reactions) = position of concentrated mass/load

F = qdx

l

F=

ke =

F

i

f ( t ) = q(t)(x )dx + Fi i

l i

k kl

The natural period of vibration for the equivalent system in the linear resistance domain is given by

T = 2 m k = 2 K l m, u M u + K l m,c M c ke

(6.7)

k = EI , xx (x ) dx

2

l

k=0

k = N , x (x ) dx

2 l

The response, y(t), is - in addition to the load history - entirely governed by the total mass, load-mass factor and the characteristic stiffness. For a linear system, the load mass factor and the characteristic stiffness are constant k = k1. The response is then alternatively governed by the eigenperiod and the characteristic stiffness. For a non-linear system, the load-mass factor and the characteristic stiffness depend on the response (deformations). Nonlinear systems may often conveniently be approximated by equivalent bi-linear or tri-linear systems, see Section 6.8. In such cases the response can be expressed in terms of (see Figure 6-6 for definitions): k1 = characteristic stiffness in the initial, linear resistance domain yel = displacement at the end of the initial, linear resistance domain T = eigenperiod in the initial, linear resistance domain and, if relevant, k3 = normalised characteristic resistance in the third linear resistance domain.

m Mi q Fi xi

i = (x = x i )

The equilibrium equation can alternatively be expressed as:

& + K(y)y = F(t) (K l m, u M u + K l m,c M c )& y

(6.6)

where

K l m,u = K l m,u = K m,u Kl K m,u Kl

= load-mass transformation factor for uniform mass transformation factor = load-mass for concentrated mass

2

Characteristic stiffness is given explicitly or can be derived from load-deformation relationships given in Section 6.10. If the response is governed by different mechanical behaviour relevant characteristic stiffness must be calculated. For a given explosion load history the maximum displacement, ymax, is found by analytical or numerical integration of equation (6.6). For standard load histories and standard resistance curves maximum displacements can be presented in design charts. Figure 6-2 shows the normalised maximum displacement of a SDOF system with a bi- (k3 = 0) or tri-linear (k3 > 0) resistance function, exposed to a triangular pressure pulse with zero rise time. When the duration of the pressure pulse relative to the eigenperiod in the initial, linear resistance range is known, the maximum displacement can be determined directly from the diagram as illustrated in Figure 6-2.

K m,u =

m (x)

l

dx

Mu

= mass transformation factor for uniform mass mass transformation factor for = concentrated mass

K m,c =

M

i

Mc

10

k 3=0.2k 1

F(t)

td

Figure 6-2 Maximum response a SDOF system to a triangular pressure pulse with zero rise time. Fmax / Rel = 2

Design charts for systems with bi- or tri-linear resistance curves subjected to a triangular pressure pulse with 0.5 td rise time is given in Figure 6.3. Curves for different rise times are given in Ch.8, Commentary Figure 8-15 to Figure 8-17. Baker's method The governing equations (6.1) and (6.2) for the maximum response in the impulsive domain and the quasi-static domain may also be used along with response charts for maximum displacement for different Fmax/Rel ratios to produce pressureimpulse (Fmax, I) diagrams - iso-damage curves - provided that the maximum pressure is known. The advantage of using iso-damage diagrams is that "backward" calculations are possible: The diagram is established on the basis of the resistance curve. The information may be used to screen explosion pressure histories and eliminate those that obviously lie in the admissible

domain. This will reduce the need for large complex simulation of explosion scenarios.

Transformation factors for elasticplastic-membrane deformation of beams and one-way slabs with different boundary conditions are given in Table 6-2. Maximum displacement for a SDOF system exposed to a triangular pressure pulse with rise time of 0.5td is displayed in Figure 6.3. Maximum displacement for a SDOF system exposed to different pressure pulses are given in Ch.8, Commentary Figure 8-15 to Figure 8-17. The characteristic response of the system is based upon the resistance in the linear range, k = k1, where the equivalent stiffness is determined from the elastic solution to the actual system.

100

Rel/Fmax=0.05

=0.1

= 0.3

= 0.5

= 0.6 = 0.7

Rel/Fmax= 0.8

10

ymax/yel

= 0.9

F

k3 = 0 k 3 = 0.1k1 k 3 = 0.2k 1 k 3 = 0.5k 1 0.1 0.1 1

R Rel

k3 = 0.5k1 =0.2k1

=0.1k1

Fmax td

k1 0.50td

yel

y

10

td/T

Figure 6-3 Dynamic response of a SDOF system to a triangular load (rise time = 0.50 td)

SDOF analysis of built-up structures (e.g. stiffeners supported by girders) is admissible if the fundamental periods of elastic vibration are sufficiently separated the response of the component with the smallest eigenperiod does not enter the elastic-plastic domain so that the period is drastically increased If these conditions are not met, then significant interaction between the different vibration modes is anticipated and a multi degree of freedom analysis is required with simultaneous time integration of the coupled system.

6.7.1 Cross-sectional behaviour

Elasto-plastic : The effect of partial yielding on bending moment is accounted for Elastic-perfectly plastic: Linear elastic up to fully plastic bending moment The simple models described herein assume elastic-perfectly plastic material behaviour. Note: Even if the analysis is based upon elastic-perfectly plastic behaviour, the material has to exhibit strain hardening in practice for the theory to be valid. The effect of strain hardening on the plastic, cross-sectional resistances may be accounted for by using an equivalent (increased) yield stress. If this is considered relevant, and the material is utilised beyond ultimate strain, it is often justified to use an equivalent yield stress equal to the mean of the lower yield stress and the ultimate stress. In the clauses for the ductility limits the effect of strain hardening is accounted for.

Moment

Component behaviour

R k1 w Elastic k2 k1 w Elastic-plastic (determinate) R k1 w Elastic-plastic (indeterminate) Elastic-plastic with membrane k2 R

k2 k1

k3

Elastic: Elastic material, small deformations Elastic-plastic (determinate): Elastic-perfectly plastic material. Statically determinate system. Bending mechanism fully developed with occurrence of first plastic hinge(s)/yield lines. No axial restraint. Elastic-plastic (indeterminate): Elastic perfectly plastic material. Statically indeterminate system. Bending mechanism develops with sequential formation of plastic hinges/yield lines. No axial restraint. For simplified analysis this system may be modelled as an elastic-plastic (determinate) system with equivalent initial stiffness. In lieu of more accurate analysis the equivalent stiffness should be determined such that the area under the resistance curve is preserved. Elastic-plastic with membrane: Elastic-perfectly plastic material. Statically determinate (or indeterminate). Ends restrained against axial displacement. Increase in load-carrying capacity caused by development of membrane forces.

Elastic + =

Rigid-plastic

6.9.1 Elastic - rigid plastic relationships In lieu of more accurate calculations rigid plastic theory combined with elastic theory may be used. In the elastic range the stiffness and fundamental period of vibration of a clamped plate under uniform lateral pressure can be expressed as:

The resistance curves in 6.7 are idealised. For simplified SDOF analysis the resistance characteristics of a real non-linear system may be approximately modelled. An example with a tri-linear approximation is illustrated in Figure 6-6. The stiffness in the k3 phase may have some contribution from strain hardening, but in most cases the predominant effect is development of membrane forces when member ends are restrained form inward displacement.

r = k1w k1 =

T= 2

ts 4 D

D=E

k3 k 2=0 k1

wel w

12 1 2

t3

800 700 600 40 35

Rel

30 25 20 15 10 5 0

In lieu of more accurate analysis the resistance curve of elasticplastic systems may be composed by an elastic resistance and a rigid-plastic resistance as illustrated in Figure 6-7.

l /s

In the plastic range the resistance (r) of plates with edges fully restrained against inward displacement and subjected to uniform pressure can be taken as:

2 + (3 2 ) r = 1+ w 9 3 rc 2

w 1

(6.8)

Clamped ends: 12f t 2 w w= rc = 2 y2 t l

2 s s s = 3+ l l l

w >1

l (>s) s t rc

w

= = = =

plate length plate width plate thickness plastic resistance in bending for plates with no axial restraint = non-dimensional displacement parameter

6 5

The effect of flexibility may be taken into account in an approximate way by means of plate strip theory and the procedure described in Section 3.7.2. The relative reduction of the plates plastic resistance, with respect to the values given in Equation (6.8), is taken equal to the relative reduction of the resistance for a beam with rectangular cross-section (plate thickness x unit width) and length equal to stiffener spacing, using the diagram for = 2 (Figure 6-12). The elastic straining of the plate is accounted for by the 2nd term in Equation (6.8). In lieu of more accurate calculation, the effect of pull-in, given by the first term in Equation (6.8) may be estimated by removing the plate and apply a uniformly distributed unit in-plane force normal to the plate edges. The axial stiffness should be taken as the inverse of the maximum in-plane displacement of the long edge. In lieu of more accurate calculation, it should be conservatively assumed that no membrane effects exist for a plate located close to an unsupported boundary, i.e. the resistance should be taken as constant and equal to the resistance in bending, r = rc over the allowable displacement range. In lieu of more accurate calculations, it is suggested to assess the relative reduction of the resistance for a uniformly loaded plate located some distance from an unsupported boundary with c = 1.0. If membrane forces are taken into account it must be verified that the adjacent structure is strong enough to anchor the fully plastic membrane tension forces.

6.9.3 Tensile fracture of yield hinges In lieu of more accurate calculations the procedure described in Section 3.10.4 may be used for a beam with rectangular cross-section (plate thickness x unit width) and length equal to stiffener spacing.

l/s = 100 5 3 2 1

Provided that the stiffeners/girders remain stable against local buckling, tripping or lateral torsional buckling stiffened plates/ girders may be treated as beams. Simple elastic-plastic methods of analysis are generally applicable. Special considerations shall be given to the effect of: Elastic flexibility of member/adjacent structure Local deformation of cross-section Local buckling Strength of connections Strength of adjacent structure Fracture

Resistance [r/rc]

Relative displacement

In Equation (6.8) the beneficial effect of membrane stiffening is represented by the term containing the non-dimensional displacement parameter w . Great caution should be exercised when assuming the presence of the membrane effect, because the membrane forces must be anchored in the adjacent structure. For plates located in the middle of a continuous plate field, the boundaries have often considerable restraint against pull-in. If the plate is located close to the boundary, the edges are often not sufficiently stiffened to prevent pull-in of edges. Unlike stiffeners no simple method with a clear physical interpretation exists to quantify the effect of flexibility on the resistance of plates under uniform pressure. In the formulations used in this RP the flexibility may be split into two contributions 1) Pull-in of edges 2) Elastic straining of the plate

6.10.1 Beams with no- or full axial restraint Equivalent springs and transformation factors for load and mass for various idealised elasto-plastic systems are shown in Table 6-2. For more than two concentrated loads, equal in magnitude and spacing, use values for uniform loading. Shear deformation may have a significant impact on the elastic flexibility and eigenperiod of beams and girders with a short span/web height ratio (L/hw), notably for clamped ends. The eigenperiod and stiffness in the linear domain including shear deformation may be calculated as:

T = 2 m k = 2 Kl m,u Mu + Kl m,c Mc

' k1

rg E A (6.9) 1+ cs L 1+ G A w

and

1 1 1 = + , ' k k k1 1 Q k Q = cQ GA w L

(6.10)

= 1.0 for both ends simply supported = 1.25 for one end clamped and one end simply supported = 1.5 for both ends clamped L = length of beam/girder E = elastic modulus G = shear modulus A = total cross-sectional area of beam/girder Aw = shear area of beam/girder kQ = shear stiffness for beam/girder k1 = bending stiffness of beam/girder in the linear domain according to Table 6-2 rg = radius of gyration

cs

Mps = plastic bending capacity of beam at support Mpm = plastic bending capacity of beam at midspan and regardless of rotational boundary conditions the following values may be used cQ = 8 for uniformly distributed loads = 4 for one concentrated loads = 6 for two concentrated loads The dynamic reactions according to Table 6-2 become increasingly inaccurate for loads with short duration and/or high magnitudes.

Table 6-2 Transformation factors for beams with various boundary and load conditions Mass factor Load-mass factor Load case Km Klm Load Maximum Resistance Factor ConcenUniConcen- Uniform resistance domain Kl Rel trated form trated mass mass mass mass

Linear stiffness k1

384 EI 5 L3

Dynamic reaction V

0.39 R + 011 . F 0.38 Rel + 012 . F

2 N P ymax L

0.78 R 0.28 F 0.75Rel 0.25F

Elastic

0.64 0.50 0.50 1.0 1.0 1.0 0.87 1.0 1.0 1.0 1.0 1.0 0.76 1.0 1.0

0.50 0.33 0.33 0.49 0.33 0.33 0.52 0.56 0.56 1.0 1.0 1.0 0.87 1.0 1.0

8Mp

F=pL

Plastic bending

L

8Mp L

0

4NP L

4Mp

48 EI

F

L/2 L/2

L

4Mp

L3

4NP L

2 N P ymax L

6Mp

L

56.4 EI

L3

F/2

F/2

Plastic bending

L/3 L/3 L/3

6Mp

L

0

6N P L

Plastic membrane

Load case

Resistance domain

12 M ps

L

8 M ps + M Pm L

384 EI

L3

Elastic

0.53

0.41

0.77

F=pL L

Elastoplastic bending Plastic bending Plastic membrane 0.64 0.50 0.78

0.36 R + 0.14 F

384 EI 5 L3

0.50 0.50

0.33 0.33

0.66 0.66

8 M ps + M Pm L

0

4NP L

L

Load case

Resistance domain

4 M ps + M Pm

192 EI

L3

48EI 3 m1 L

Elastic

1.0 1.0 1.0 080 0.87 1.0 1.0 0.58 0.64 0.50 0.50 1.0 1.0 1.0 1.0 0.81 0.87 1.0 1.0

0.37 0.33 0.33 0.41 0.52 0.56 0.56 0.45 0.50 0.33

0.37 0.33 0.33 0.51 0.60 0.56 0.56 0.78 0.78 0.66

( (

) )

0.71R 0.21F

F

L/2 L/2

L

4 M ps + M Pm

Plastic bending Plastic membrane Elastic Elastoplastic bending Plastic bending Plastic membrane Elastic

0

4NP L

F/2

F/2

9 M ps L

6 ( M ps + M Pm ) L

6 ( M ps + M Pm ) L

260 EI L3

56.4 EI L3

212 EI m1 3 L

0.48R + 0.02 F

0.52 Rel 0.02 F

0

6N P L

8 M ps

L

4 M ps + 2 M Pm L

4 M ps + 2 M Pm L

185EI

L

3

160 EI m2 3 L

F=pL

V1

Elastoplastic bending

V2

)

)

384 EI 5L

3

V1 = 0.25R + 0.07 F

V2 = 0.54 R + 014 . F

0

4NP L

0.33 1.0 1.0 1.0 1.0 0.67 0.76 1.0 1.0 0.43 0.49 0.33 0.33 0.45 0.52 0.56 0.56

0.66 1.0 1.0 1.0 1.0 0.83 0.87 1.0 1.0 0.43 0.49 0.33 0.33 0.55 0.60 0.56 0.56

6 M Ps L

2 M ps + 3 M Pm L

16 M Ps 3L

2 M ps + 2 M Pm L

107 EI

L3

F

V1 L/2 L/2 V2

48 EI

L3

160 EI m2 3 L

V1 = 017 . R + 017 . F V2 = 0.33R + 0.33F

2 M ps + 2 M Pm L

0

4NP L 132 EI L3

F/2 V1

F/2 V2

56 EI L3

122 EI m3 3 L

2 M ps + 3 M Pm L

0

6N P L

Where: q = explosion load per unit length = ps for stiffeners = p l for girders

m1 =

m2 = m3 =

1.5M ps

M ps + M pm

1.5M ps

M ps + 2M pm

2M ps

M ps + 3M pm

6.10.2 Beams with partial end restraint. Relatively small axial displacements have a significant influence on the development of tensile forces in members undergoing large lateral deformations. Equivalent elastic, axial stiffness may be defined as

w=

w c1 w c

= non-dimensional deformation characteristic beam height for beams = described by plastic interaction equation (6.12). = non-dimensional spring stiffness

l 1 1 = + k k node 2EA

(6.11)

wc = c=

Wp A

2

knode = axial stiffness of the node with the considered member removed. This may be determined by introducing unit loads in member axis direction at the end nodes with the member removed. Plastic force-deformation relationship for a beam under uniform pressure may be obtained from Figure 6-10, Figure 6-11 or Figure 6-12 if the plastic interaction between axial force and bending moment can be approximated by the following equation:

4c1 kw c f y Al

c1 = 2 c1 = 1 WP

Wp = zgAg A = As + st Ae = As + set zg As s se

M N + Mp N p

=1

(6.12)

In lieu of more accurate analysis = 1.2 can be assumed for stiffened plates and H or I beams. For members with tubular section = 1.5. For rectangular sections and plates = 2.0 can be assumed.

R0 = l

8c 1 f y Wp l

plastic collapse resistance in bending for = the member with uniform load. = member length

= for clamped beams = for pinned beams = plastic section modulus for the cross section of the beam = plastic section modulus for stiffened plate for set > As = total area of stiffener and plate flange = effective cross-sectional area of stiffener and plate flange, = distance from plate flange to stiffener centre of gravity. = stiffener area = stiffener spacing = effective width of plate flange see Section 6.10.4

= 1.2

R/R 0

F

0.5 1 0 0.2 0.1

k

(explosion load)

c=

Deformation

Figure 6-10 Plastic load-deformation relationship for beam with axial flexibility ( = 1.2)

6 5

= 1.5

R/R0

(explosion load)

3 2

c=

0.5

0.2

0.1

w

0

Deformation

Figure 6-11 Plastic load-deformation relationship for beam with axial flexibility ( = 1.5)

9 8 7 6

=2

Bending & membrane Membrane only

R/R0

5 4 3 2

0.2 0.5

0.1

F

k

(explosion load)

c=

1 0

Deformation

Figure 6-12 Plastic load-deformation relationship for beam with axial flexibility ( = 2)

For members where the plastic moment capacity of adjacent members is smaller than the moment capacity of the exposed member the force-deformation relationship may be interpolated from the curves for pinned ends and clamped ends:

where

0=

R0 8

actual

R = R clamped + (1 )R pinned

(6.13)

Mp l

1 1

(6.14)

actual 0

= Collapse load in bending for beam accounting for actual bending resistance of adjacent members

Commentary. Determination of effective flange due to buckling can be made as for buckling of stiffened plates see DNV-RP-C201. The effective width for elastic deformations may be used when the plate flange is on the tension side. In most cases the flange will experience varying stress with parts in compression and parts in tension. It may be unduly conservative to use the effective width for the section with the largest compression stress to be valid for the whole member length. For continuous stiffeners it will be reasonable to use the mean value between effective width at the section with the largest compression stress and the full width. For simple supported stiffeners with compression in the plate it is judged to be reasonable to use the effective width at midspan for the total length of the stiffener.

6.10.5 Strength of adjacent structure

R 0actual =

i

8M p + 4M p1 + 4M p2 l

(6.15) (6.16)

M Pj = M Pj,i M P

i = adjacent member no i j = end number {1,2} MPj,i = Plastic bending moment for member no. i. Elastic, rotational flexibility of the node is normally of moderate significance

6.10.3 Beams with partial end restraint - support capacity smaller than plastic bending moment of member

For beams where the plastic moment capacity of adjacent members is smaller than the moment capacity of the impacted beam, the force-deformation relationship, R*, may be derived from the resistance curve, R, for beams where the plastic moment capacity of adjacent members is larger than the moment capacity of the impacted beam (Section 3.7.2), using the expression:

R * = R + (R 0 R * 0) R* = R

w w lim

,

The adjacent structure must be checked to see whether it can provide the support required by the assumed collapse mechanism for the member/sub-structure

6.10.6 Strength of connections

The capacity of connections can be taken from recognised codes. The connection shall be checked for the dynamic reaction force given in Table 6-2. For beams with axial restraint the connection should also be checked for lateral - and axial reaction in the membrane phase: If the axial force in a tension member exceeds the axial capacity of the connection the member should be assumed disconnected. If the capacity of the connection is exceeded in compression and bending, this does not necessarily mean failure of the member. The post-collapse strength of the connection may be taken into account provided that such information is available.

6.10.7 Ductility limits

(6.17)

1.0

where R0 = Plastic bending resistance with clamped ends (c1 = 2) moment capacity of adjacent members larger than the plastic bending moment of the beam = Plastic bending resistance - moment capacity of adjaR* 0 cent members at one or both ends smaller than the plastic bending moment of the member

R* 0 =

M Pj = M Pj,i M P

i

4M P + 2M P1 + 2M P2 l

(6.18) (6.19)

Reference is made to Section 3.10. The local buckling criterion in Section 3.10.2 and tensile fracture criterion given in Section 3.10.3 may be used with: dc c = characteristic dimension equal to twice the distance from the plastic neutral axis in bending to the extreme fibre of the cross-section = non-dimensional axial spring stiffness calculated in Section 6.10.2.

y

i j MPj,i wlim

= = = =

adjacent member no i end number {1,2} Plastic bending resistance for member no. i. limiting non-dimensional deformation where the membrane force attains yield, i.e. the resistance curve, R, with actual spring stiffness coefficient, c, intersects with the curve for c = . If c = ,

wlim =

Table 6-3 Ductility ratios - beams with no axial restraint Boundary conditions Load Cross-section type 1) Type I Type II Type III

6.10.4 Effective flange

In order to analyse stiffened plate as a beam the effective width of the plate between stiffeners need to be determined. The effective width needs to be reduced due to buckling and/or shear lag. Shear lag effects may be neglected if the length is more than 2.5 times the width between stiffeners. For guidance see Ch.8,

1)

6 7 6 12 6 4

4 5 4 8 4 3

2 2 2 3 2 2

7. References

/1/ /2/ NORSOK Standard N-003 Action and Action Effect NS-ENV 1993-1 Eurocode 3: Design of Steel structures Part 1-2. General rules - Structural fire design /3/ Amdahl, J.: Energy Absorption in Ship-Platform Impacts, UR-83-34, Dept. Marine Structures, Norwegian Institute of Technology, Trondheim, 1983. /4/ SCI 1993: Interim Guidance Notes for the Design and Protection of Topside Structures against Explosion and Fire /5/ Amdahl, J.: Mechanics of Ship-Ship Collisions- Basic Crushing Mechanics. West Europene Graduate School of Marine Technology, WEGEMT , Copenhagen, 1995 /6/ Skallerud, B. and Amdahl, j.: Nonlinear Analysis of Offshore Structures, Research studies Press, UK 2002 /7/ Amdahl, J. and Johansen, A.: High-Energy Ship Collision with Jacket Legs ISOPE, Stavanger, 2001 /8/ Moan, T., Amdahl, J., Wang, X. and Spencer, J.: Risk Assessment of FPSOs, with Emphasis on Collisions, SNAME Annual Meeting, Boston, 2002 /9/ Skallerud, B. and Amdahl, j.: Nonlinear Analysis of Offshore Structures, Research studies Press, UK 2002 /10/ Amdahl, J. and Johansen, A.: High-Energy Ship Collision with Jacket Legs ISOPE, Stavanger, 2001

The added mass is due to the hydrodynamic pressure induced by the forced motion of water particles on the wet surface of the ship. By solving the velocity potential for the fluid on the body surface, the added mass is determined by means of 2-D (strip theory) or 3-D techniques. The added mass is frequency dependent, and thus varies with time during a collision, but a constant value is recommended for simple analysis. The fraction of collision energy to be dissipated as strain energy for shuttle tanker impact on FPSO stern is shown in Figure 8-1. Note the strong dependency of the mass ratio; the larger the mass of shuttle tanker, the lesser of the collision energy must be dissipated as strain energy. (However, provided that the speed of the shuttle tanker is constant, the absolute value of the strain energy increases) The relative size may differ considerably for the approach phase (shuttle tanker in ballast, FPSO fully loaded) and the departure phase (shuttle tanker fully loaded, FPSO in ballast). To illustrate this, possible values are listed in Table 8-1. In this example both the FPSO and shuttle tanker are large compared to typical North Sea conditions. The same added mass coefficient applies to both vessels. It is observed that the fraction of energy to be dissipated as strain energy varies between 0.33 (departure) and 0.71 (approach). This indicates that the approach phase may be particularly critical with respect to the consequences of collision.

Table 8-1 Fraction of collision energy for strain energy dissipation Vessel size [dwt] Approach phase Departure phase Shuttle tanker 150.000 370.000 FPSO 320.000 160.000 Strain energy fraction 0.71 0.33

8. Commentary

Comm. 2.3 General The structural design is seen as having acceptable safety against accidental loads when the design accidental loads are less than the design resistance. This is similar to the check of the structure for ordinary loads but with the following differences: As ordinary loads are either permanent or occur frequent it will not be acceptable that the load lead to reduced load carrying capacity while the short duration and the low probability of accidental loads make this an acceptable assumptions. E.g. a blast wall need not be capable of resisting another explosion after a blast, but if the wall is used as a fire barrier it need to serve as such after the blast. Comm. 3.1 General For typical installations, the contribution to energy dissipation from elastic deformation of component/substructures in direct contact with the ship is very small and can normally be neglected. Consequently, plastic methods of analysis apply. However, elastic elongation of the hit member as well as axial flexibility of the nodes to which the member is connected, have a significant impact on the development of membrane forces in the member. This effect has to be taken into account in the analysis, which is otherwise based on plastic methods. The diagrams in Section 3.7.2 are based on such an approach. Depending on the structure size/configuration as well as the location of impact elastic strain energy over the entire structure may contribute significantly. Comm. 3.2 Design principles The transition from essentially strength behaviour to ductile response can be very abrupt and sensitive to minor changes in scantlings. E.g. integrated analyses of impact between the stern of a supply vessel and a large diameter column have shown that with moderate change of (ring - and longitudinal) stiffener size and/or spacing, the energy dissipation may shift from predominantly platform based to predominantly vessel based. Due attention should be paid to this sensitivity when the calculation procedure described in Section 3.5 is applied. Comm. 3.3 Collision mechanics

1,2 Strain energy fraction 1 0,8 0,6 0,4 0,2 0 0 1 2 3 4 Mass ratio [(ms + as)/(mi+ ai)]

Figure 8-1 Fraction of collision energy to be dissipated as strain energy for shuttle tanker impact on an FPSO.

The force-deformation relationship for impacts from supply vessels/merchant vessels against jacket legs have been elaborated because of the need to consider high-energy impacts (collision energy ~50 MJ) for some installations in the North Sea. The likelihood of a central impact against a leg is obviously not very large, but has still been considered because loss of a leg could be critical for some platforms. Experience has shown, however, that many large North Sea jackets have sufficient strength to crush the bow. Reference is made to Amdahl and Johansen (2001).

The curve for bow impact in Figure 3-4 has been derived on the assumption of impacts against an infinitely rigid wall. Sometimes the curve has been used erroneously to assess the energy dissipation in bow-brace impacts. Experience from small-scale tests Ch.7, /3/ indicates that the bow undergoes very little deformation until the brace becomes strong enough to crush the bow. Hence, the brace absorbs most of the energy. When the brace is strong enough to crush the bow the situation is reversed; the brace remains virtually undamaged. On the basis of the tests results and simple plastic methods of analysis, force-deformation curves for bows subjected to (strong) brace impact were established in Ch.7, /3/ as a function of impact location and brace diameter. These curves are reproduced in Figure 8-2. In order to fulfil a strength design requirement the brace should at least resist the load level indicated by the broken line (recommended design curve). For braces with a diameter to thickness ratio < 40 it should be sufficient to verify that the plastic collapse load in bending for the brace is larger than the required level. For larger diameter to thickness ratios local denting must probably be taken into account. Normally sized jacket braces are not strong enough to resist the likely bow forces given in Figure 8-2, and therefore it has to be assumed to absorb the entire strain energy. For the same reasons it has also to be assumed that the brace has to absorb all energy for stern and beam impact with supply vessels.

Impact force [MN] Recommended design curve for brace impact 12 Between a stringer (D= 1.0 m) 8 On a stringer (D= 0.75 m) Between stringers (D= 0.75) m

Figure 8-3 shows an example of the evolution of contact force intensity during a collision between the stern corner of a supply vessel and a stiffened column. In the beginning the contact is concentrated at the extreme end of the corner, but as the corner deforms it undergoes inversion and the contact ceases in the central part. The contact area is then, roughly speaking, bounded by two concentric circles, but the distribution is uneven. The force-deformation curves given in Figure 3-4 relate to total collision force for stern end - and stern corner impact , respectively. Table 3-1 and Table 3-2 give the anticipated maximum force intensities (local force and local contact areas, i.e. subsets of the total force and total area) at various stages of deformation. The basis for the design curves is integrated, non-linear finite element analysis of stern/column impacts. The information given in 3.5.2 may be used to perform strength design. If strength design is not achieved numerical analyses have shown that the column is likely to undergo severe deformations and absorb a major part of the strain energy. In lieu of more accurate calculations (e.g. non-linear FEM) it has to be assumed that the column absorbs all strain energy.

Comm. 3.5.3 Energy dissipation is ship bow.

The requirements in this paragraph are based upon considerations of the relative resistance of a tubular brace to local denting and the bow to penetration of a tubular beam. A fundamental requirement for penetration of the brace into the bow is, first - the brace has sufficient resistance in bending, second - the cross-section does not undergo substantial local deformation. If the brace is subjected to local denting, i.e. undergoes flattening, the contact area with the bow increases and the bow inevitably gets increased resistance to indentation. The provisions ensure that both requirements are complied with. Figure 8-8 indicates the level of the various contact locations. Figure 8-4 shows the minimum thickness as a function of brace diameter and resistance level in order to achieve sufficient resistance to penetrate the ship bow without local denting. It may seem strange that the required thickness becomes smaller for increasing diameter, but the brace strength, globally as well as locally, decreases with decreasing diameter. Local denting in the bending phase can be disregarded provided that the following relationship holds true:

D 1l 0.14 2 t c1 D

2

0.5

1.0

Comm. 3.5.2 Force contact area for strength design of large diameter columns.

(8.1)

Area with high force intensity Deformed stern corner

Figure 8-5 shows brace thickness as a function of diameter and length diameter ratio that results from Equation (8.1). The thickness can generally be smaller than the values shown, and still energy dissipation in the bow may be taken into account, but if Equation (8.1) is complied with denting does not need to be further considered. The requirements are based upon simulation with LS-DYNA for penetration of a tube with diameter 1.0 m. Great caution should therefore be exercised in extrapolation to diameters substantially larger than 1.0 m, because the resistance of the bow will increase. For brace diameters smaller than 1.0 m, the requirement is conservative.

Figure 8-3 Distribution of contact force for stern corner/large diameter column impact

fy = 235 MPa, 6 MN fy = 235 MPa, 3 MN fy = 355 MPa, 6 MN fy = 355 MPa, 3 MN fy = 420 MPa, 6 MN fy = 420 MPa, 3 MN

Comm. 3.10.1 General

If the degradation of bending capacity of the beam cross-section after buckling is known the load-carrying capacity may be interpolated from the curves with full bending capacity and no bending capacity according to the expression:

R ( w ) = RM P =1 ( w ) + RM P =0 ( w )(1 )

(8.2)

Figure 8-4 Required thickness versus grade and resistance level of brace to penetrate ship bow without local denting

R M P , red R M P =1 ( w = 0)

100 Thickness [mm] 80 60 40 20 0 0,6 0,8 1 1,2 Diameter [m] 1,4 L/D =20 L/D =25 L/D =30

R M P , red

= Plastic collapse load in bending with reduced cross-sectional capacities. This has to be updated along with the degradation of crosssectional bending capacity.

Figure 8-5 Brace thickness yielding little local deformation in the bending phase

The rupture criterion is calculated using conventional beam theory. A linear strain hardening model is adopted. For a cantilever beam subjected to a concentrated load at the end, the strain distribution along the beam can be determined from the bending moment variation. In Figure 8-7 the strain variation, = cr Y , is shown for a circular cross-section for a given hardening parameter. The extreme importance of strain hardening is evident; with no strain hardening the high strains are very localised close to the support (x = 0), with strain hardening the plastic zone expands dramatically. On the basis of the strain distribution the rotation in the plastic zone and the corresponding lateral deformation can be determined. If the beam response is affected by development of membrane forces it is assumed that the membrane strain follows the same relative distribution as the bending strain. By introducing the kinematic relationships for beam elongation, the maximum membrane strain can be calculated for a given displacement.

50 45 40 35

Comm. 3.7.3 Support capacity smaller than plastic bending moment of the beam

The procedure is illustrated in Figure 8-6. Elastic, rotational flexibility of the node is normally of moderate significance.

7 6 R/R0 R*/R0 5 4 3 2 1 0 0 1 2

R* 0 / R0

= 1.5

c =0.5

No hardening

P x l

Strain

30 25 20 15 10

R/R0

R*/R0 wlim 3 4

Deformation w

x/l

Figure 8-6 Derivation of force-deformation relationship for beam with end moments less than beam plastic moment.

Figure 8-7 Axial variation of maximum strain for a cantilever beam with circular cross-section

The procedure given is essentially the same as the one used in NORSOK N-004, but is formulated differently. The bending moment boundary condition is important in the bending phase, but has no influence on the resistance in the pure membrane

Adding the bending strain and the membrane strain allows determination of the critical displacement as a function of the total critical strain. Figure 8-8 shows deformation at rupture for a fully clamped beam as a function of the axial flexibility factor c.

5 4.5 4 3.5 3

20% 15%

l/D = 30 c= 0 = 0.05 = 0.5 = 1000 l/D = 20 c= = = = 0 0.05 0.5 1000

Strain

w/D

USFOS shell

40

60

80

100

120

cr/y

Figure 8-8 Maximum deformation for a tubular fully clamped beam (H=0.005)

The plastic stiffness factor H is determined from the stressstrain relationship for the material. The equivalent linear stiffness shall be determined such that the total area under the stress-strain curve up to the critical strain is preserved (The two portions of the shaded area shall be equal), refer Figure 8-9. It is un-conservative and not allowable to use a reduced effective yield stress and a plastic stiffness factor as illustrated in Figure 8-10.

It is observed that the strain estimated in ABAQUS analysis depend significantly on the mesh size evidencing the need for a mesh-size-dependent fracture strain criterion. The NORSOK criterion agrees fairly well with FEM calculations when a fine mesh is used. The criterion is conservative, as desired. The strain calculation in the USFOS beam element assumes a yield plateau followed by parabolic type hardening. Only the fine ABAQUS mesh captures the yield plateau effect. Comm. 3.12 Energy dissipation in floating production vessels

fcr

HE E

cr

Figure 8-9 Determination of plastic stiffness

cr

1600

1600

f

HE

The accuracy of the calculation model for tensile fracture in yield hinges has been investigated by Amdahl and Skallerud (2002). The maximum strain as a function of lateral displacement (Equation (3.22)) for a tubular beam is compared with the maximum strain from finite element calculations in Figure 8-11. The beam is assumed to be clamped and fixed against inward axial displacement, l = 25 m, D = 1 m, t = 0.06 m, fy = 300 MPa, H = 0.00287 (i.e. ultimate stress fu = 390 MPa for at ultimate strain u = 0.15). The mesh size for USFOS shell and ABAQUS is 0.25 0.39 m and for ABAQUS fine mesh 0.05 0.195 m. The element used in ABAQUS analyses is the S4R reduced integration element .

Calculation of energy dissipation in stringers, decks and bulkheads subjected to gross, axial crushing shall be based upon recognised methods for plastic analysis, e.g. Ch.7, /3/ and Ch.7, /4/. It shall be documented that the folding mechanisms assumed yield a realistic representation of the true deformation field. The force deformation relationships given in Figure 3-6 may be used to design a collision resistant stern of an FPSO. In order to be impact resistant, stringers and frames must be fairly closely spaced, typically in the range of 1.5 2 m. Given the relative dimensions of the girder system and the bulb crosssection, as illustrated in Figure 8-12, it is reasonable to apply the total collision force as uniformly distributed line loads on the stringers and frames. The integrity of the stringers and frames can then be checked in a FEM analysis. Moderate local yielding should be accepted. The stern structure must resist the collision force during all stages of the collision process. Normally, it suffices to analyse a few collision force and contact area situations. It is normally neither practical nor necessary to design the plating and stiffeners such that their response is elastic. Large plastic deformations can be accepted, but fracture of the plating should not occur (Note: provided that strength design is aimed for). In lieu of more accurate calculations, the contact force may be considered uniformly distributed over the plate field, and the resistance may be assessed using the provisions given for the resistance of plates and stiffeners to explosion loads.

1600

fcr

HE

Even if the stiffeners are allowed to deform under extreme collision loads, they should be sufficiently robust to initiate crushing of the bulb. Engineering judgment must be applied, but it is recommended to design the stiffeners according to requirements for ships navigating in ice; DNV Ice Class POLAR. With respect to deformation resistance of stiffened plating, see next paragraph. The ductile resistance of stiffened plates may be analysed considering the side as an assembly of plate/stiffeners. The resistance of individual stiffeners with associated plate flange can be calculated with the methods given in Section 6.3 using relationships for a concentrated force, see example in Ch.8, Comm. 9.3. The resistance of the various stiffeners will be mobilised according to the geometry (raking) of the impacting bow. Unless the frame spacing is long or the stiffener height is small, fracture will take place before noticeable membrane stiffening has taken place. The initiation of fracture does not necessarily

imply that the resistance is totally lost, because fracture takes place in the top flange while the strain on the plate side is considerably smaller . The above procedure neglects the effect of membrane forces transverse to the stiffeners. Depending on the geometry of the panel this contribution may be substantial. Collisions with FPSOs have been studied in-depth in a paper by Moan et.al. (2002). Force-deformation relationships are given for supply vessels/merchant vessels, 18.000 tons chemical tanker and a 42.000 tons tanker and a shuttle tanker. The collision risk for all categories of vessels is discussed extensively. The consequences of a collision with a shuttle tanker servicing the FPSO are especially considered. Figure 8-13 shows the force-deformation relationship for supply vessel/merchant vessel colliding with the side of an FPSO. It is interesting to see that the force level for bow without bulb is smaller than the bow force-deformation curve given in Figure 3-4.

30 25 Energy [MJ] 20 15 10 5 0 0

Energy s upers tr. Energy bulb Total force Forc e bulb Forc e superstr.

30 25 Force [MN] 20 15 10 5 0

Figure 8-13 Force-deformation relationship for supply vessel/merchant vessel impact against FPSO side

The validity for the energy equation 4.6 is limited to 7 < 2 r/d < 41, t/d < 0.22 and mi/m < 0.75. The formula neglect the local energy dissipation which can be added as Eloc = R0.2 t. In case of hit near the plate edges the energy dissipation will be low and may lead to unreasonable plate thickness. The failure criterion used for the formula is locking of the plate. In many cases locking may be acceptable as long as the falling object is stopped. If the design is based on a hit in the central part of a plate with use of the smaller diameter in the treaded part in the calculations, no penetration of the drill collar will take place at any other hit location due to the collar of such dropped objects.

Comm. 5.1 General

For redundant structures thermal expansion may cause buckling of members below 400C. Forces due to thermal expansion are, however, purely internal and will be released once the member buckles. The net effect of thermal expansion is therefore often to create lateral distortions in heated members. In most cases these lateral distortions will have a moderate influence on the ultimate strength of the system. As thermal expansion is the source of considerable inconvenience in conjunction with numerical analysis it would tempting

to replace its effect by equivalent, initial lateral member distortions. There is however, not sufficient information to support such a procedure at present. Comm. 5.5 Empirical correction factor In Ch.7 /2/ Eurocode 3 an empirical reduction factor of 1.2 is applied in order to obtain better fit between test results and column curve c for fire exposed compressive members. In the design check this is performed by multiplying the design axial load by 1.2. In non-linear analysis such a procedure is impractical. In non-linear space frame, stress resultants based analysis the correction factor can be included by dividing the yield compressive load and the Euler buckling load by a factor of 1.2. (The influence of axial force on members stiffness is accounted for by the so-called Liveslys stability multipliers, which are functions of the Euler buckling load.) In this way the reduction factor is applied consistently to both elastic and elasto-plastic buckling. The above correction factor comes in addition to the reduction caused by yield stress and elastic modulus degradation at elevated temperature if the reduced slenderness is larger than 0.2. Comm. 6.2 Classification of response Equation (6.2) is derived using the principle of conservation of momentum to determine the kinetic energy of the component at the end of the explosion pulse. The entire kinetic energy is

then assumed dissipated as strain energy. Equation (6.3) is based on the assumption that the explosion pressure has remained at its peak value during the entire deformation and equates the external work with the total strain energy. In general, the explosion pressure is not balanced by resistance, giving rise to inertia forces. Eventually, these inertia forces will be dissipated as strain energy. Equation (6.4) is based on the assumption that the pressure increases slowly so that the static condition (pressure balanced by resistance) applies during the entire deformation. Comm. 6.4 SDOF system analogy The displacement at the end of the initial, linear resistance domain yel will generally not coincide with the displacement at first yield. Typically, yel represents the displacement at the initiation of a plastic collapse mechanism. Hence, yel is larger than the displacement at first yield for two reasons: i) Change from elastic to plastic stress distribution over beam cross-section ii) Bending moment redistribution over the beam (redundant beams) as plastic hinges form

Figure 8-14 is derived from the dynamic response chart for a SDOF system subjected to a triangular load with zero rise time given in Figure 6-3. In the example it is assumed that from ductility considerations for the assumed mode of deformation a maximum displacement of ten times elastic limit is acceptable. Hence the line y allow y max = = 10 represents the upper limit for the y el y el displacement of the component. From the diagram it is seen that several combinations of pulses characterised by Fmax and td may produce this displacement limit. Each intersection with a response curve (e.g. k3 = 0) yields a normalized pressure

F Fmax = R R el

and a normalised impulse

1 Fmax t d I = 2 = RT R el T

t 1 d R el T 2 Fmax

11 10 9 8

Pressure F/R

7 6 5 4 3 2 1 0 0

Impulsive asymptote

By plotting corresponding values of normalised impulse and normalised pressure the iso-damage curve given in Figure 8-14 is obtained. If the displacement shape function changes as a non-linear structure undergoes deformation the transformation factors change. In lieu of accurate analysis an average value of the combined load-mass transformation factor can be used:.

k lm

average

k lm

elastic

+ ( 1)k lm

plastic

(8.3)

= ymax/yel ductility ratio Since is not known a priori iterative calculations may be necessary. Dynamic response charts for a SDOF system with triangular pressure pulses with rise time different from td/2 are given in Figure 8-15 to Figure 8-17.

8 9 10 11

Pressure asymptote

1 2 3 4 5 6 7

Impulse I/(RT)

Rel/Fmax=0.05

100

=0.1

= 0.3

= 0.5

= 0.6

= 0.9

10

= 1.0

ymax/yel

F Fmax

R Rel

k3 = 0.5k1 =0.2k1

=0.1k1

k1

td

0.1 0.1 1

yel

y

10

td/T

Figure 8-15 Dynamic response of a SDOF system to a triangular load (rise time=0)

Rel/Fmax=0.05

100

=0.1

= 0.3

= 0.5

= 0.6

= 0.7

Rel/Fmax= 0.8

10

= 0.9

ymax/yel

k3 = 0.5k1 =0.2k1 =0.1k1

F Fmax

R Rel

k1 0.15td

td

yel

y

10

td/T

Figure 8-16 Dynamic response of a SDOF system to a triangular load (rise time = 0.15td)

100

Rel/Fmax=0.05

=0.1

= 0.3

= 0.5

= 0.6

= 0.7

Rel/Fmax= 0.8

10

ymax/yel

= 0.9

F

k3 = 0 k 3 = 0.1k1 k 3 = 0.2k 1 k 3 = 0.5k 1 0.1 0.1 1

R Rel

k3 = 0.5k1 =0.2k1

=0.1k1

Fmax td

k1 0.30td

yel

y

10

td/T

Figure 8-17 Dynamic response of a SDOF system to a triangular load (rise time = 0.30td)

Comm.6.7.1.1 Component behaviour For beams the characteristic linear stiffness given for the elasto-plastic resistance domain in Table 6-2 is derived from the equal area principle on the assumption that the support moment is equal to the plastic bending moment of the beam. Comm. 6.7.1.1 Component behaviour For deformations in the elastic range the effective width (shear lag effect) of the plate flange, se, of simply supported or clamped stiffeners/girders may be taken from Figure 8-18.

1.2 1 0.8

Comm. 6.10.7 Ductility limits The table is taken from Ch.7, Reference /4/. The values are based upon a limiting strain, elasto-plastic material and crosssectional shape factor 1.12 for beams and 1.5 for plates. Strain hardening and any membrane effect will increase the effective ductility ratio. The values are likely to be conservative.

9. Examples

9.1 Design against ship collisions

9.1.1 Jacket subjected to supply vessel impact The location of contact is at brace mid-span and the force acts parallel to global x-axis. The brace dimensions are 762 x 28.6 mm. From linear elastic analysis it is found that the stiffness of nodes 508 and 628 against displacement in the brace direction is 736 MN/m and 51 MN/m respectively, when the brace is removed. The unequal stiffness may be represented by two equal springs, each with stiffness:

Uniform distribution or

n> n= n= n=

6 5 4 3

s e/s

= 0.6L

/s

Figure 8-18 Effective flange for stiffeners and girders in the elastic range

It must also be verified that the capacity of the joints is sufficient to support the force state in the brace both in the bending mode of deformation and in the membrane tension state. Figure 9-3 displays the simulated bending moment-axial force interaction history in the brace and shows that the membrane force becomes substantial, but doe not attain the fully plastic axial force. In lieu of accurate calculations, it should be assume that the fully plastic tension is developed. Provided that the joints and adjacent structure are capable of supporting the brace ends, the energy dissipation is limited by fracture due to excessive straining of the brace. Fracture criteria are given Section 3.10.3. Using the fracture criterion in Section 3.10.3 there is obtained wcrit = 2.2 m and a corresponding energy dissipation E = 6 MJ.

628

508 762 x 28.6 mm l= 23.3 m

10 8 6 USFOS 4 Simple model 2 0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Displacement [m] Energy dissipation [MJ] Energy dissipation

8 6 4 2 0

1 1 K node = 2 + 736 51

= 95 MN / m

2 EA 2 2.1 10 0.762 0.0286 = = 1234 MN / m l 23.3

5

and is large compared to the stiffness of the node. This yields an effective stiffness of

1 1 1 = + = 88 MN/m K 95 1234

1.0 Normalised force N/NP 0.8 0.6 0.4 0.2 0.0 -0.2 0.0 0.2 0.4 0.6 0.8 Normalised moment M/MP 1.0

Assuming clamped ends (c1 = 2) the non-dimensional spring stiffness comes out to be

4c 1 Kw c c= f y Al

2

The resulting end restraint is quite flexible. This is particularly due to low stiffness in node 628, in spite of the support by the adjacent braces. Hence, the build-up of tension force will be delayed compared to a full axial fixity. The collapse load in bending is calculated assuming clamped conditions at both ends. This is a good approximation at the lower end but slightly optimistic at the upper end.

4c1M P 4 2 355 (0.762 0.0286) 2 0.0286 = = 1.9 MN l 23.3

R0 =

The load-deformation characteristics for the brace are obtained by interpolation of the curves given in Figure 3-7. The result is depicted in Figure 9-2. The response predicted by means of the nonlinear analysis program USFOS is also plotted. It appears that the simplified approach performs very well when axial flexibility is taken into account. The loss of stiffness predicted by USFOS at large displacements is due to initiation of failure of adjacent members at node 628. Collapse of these members takes place at a load level of 2.8 MN.

Tensile fracture of the brace considered in is estimated. The characteristic dimension is, dc = D = 0.762 m. For steel grade S 355 a strain hardening coefficient of H = 0.0034 is used, refer Table 3-3. c1 = 2 (clamped ends are assumed), the collision occurs at mid span, hence = 0.5, and l /dc = 15.3. The nondimensional spring stiffness is c = 0.18 and W/WP = /4. This yields wcrit = 2.2 m. Because of the large l /dc ratio, the brace is capable of deforming almost three times its diameter.

9.2.1 Geometry The geometry of the structure is outlined in Figure 9-4. The plate, stiffeners and girders will be assessed. The main dimensions are:

t s l

= 10 mm = 500 mm = 2000 mm

eigenperiod is adjusted by Tmod = T 1 0.65 = 5.0 msecs to account for less stiffness. This yields td/Tmod = 4.0. From the response chart there is obtained ymax/yel ~ 0.7. This yields wmax = 0.7 47,3 = 33 mm. All these methods yield approximately the same result. The static approach is quite good, but there is a slight dynamic amplification > 1 in the present case. The plate must be checked with respect to rupture, see Section 9.4.1. It is noticed that if no membrane force can be taken into account, i.e. c = 0, then ymax/yel >> 100 and the plate will fail completely.

9.2.3 Calculation of dynamic response of stiffened plate The dynamic response of the stiffened plate considered in Section 9.3.2, subjected to a triangular explosion pulse with duration 20 msecs and peak pressure fmax = 2.5 MPa, is studied. The collapse resistance is R0 = Rel = 0.58 MN, and no membrane stiffening can be assumed, i.e. k3 = 0. As the plate/stiffener undergoes a phase with elasto-plastic bending, the resistance is approximated by a linear elastic-perfectly plastic model, with equivalent stiffness of 208 MN/m and wel = 2.8 mm. The critical deformation at rupture wcrit = 36 mm, hence the ductility ratio is = ymax/yel = 36/2.8 = 13. The total mass is 108 kg. The load-mass factor is ~ 0.77 and 0.66 in the elastic/elasto-plastic and plastic bending phase, respectively. Using Equation (8.3) the average load-mass facaverage tor becomes klm = ( 0.77 + (13 1) 0.66 ) /13 = 0.67 and the eigenperiod is:

k lm

average

Girder

t = 10

Stiffener Hp180

9.2.2 Calculation of dynamic response of plate: The dynamic response of the plate considered in Section 9.3.1 is studied. The plate is subjected to a triangular pressure pulse with duration of 20 msecs. The peak pressure is fmax = 2.5 MPa. Assume that the resistance curve for c = 1.0 in Figure 97 applies. This yields rel/fmax = 0.3. The curve is redrawn below along with approximate relationships

5

P late c = 1.0

T = 2

k1

= 3.7m sec s

Resistance [Mpa]

4 3 2 1 0 0 10

S tatic

This gives td/T = 5.4. By inspection of Figure 6-3 it is found that for = 13 and td/T =5.4 Rel/Fmax 0.75 (in other words, because of limited pulse duration it is possible to overload the stiffener by 33% compared to the static collapse resistance in bending). The maximum peak pressure the stiffener can resist is:

f crit =

20 30 D eform ation [m m ]

40

50

Alternative 1- static analysis: The eigenperiod of the plate according to Section 9.3.1 with = 25 is T = 4.0 msecs. Hence td/T = 5. This is a fairly long duration and static behaviour can be assumed. The maximum deflection is determined directly from Figure 9-7, i.e. wmax = 27 mm. Alternative 2 - tri-linear resistance: By inspection of the dynamic response charts and the resistance curve for the plate it is noticed that none of the tri-linear curves apply very well. The best fit is obtained with k3 = 0.5 k1, but this underestimates the resistance for large deformations. From the response chart for td/T = 5 there is read ymax/yel ~ 4.8. This yields wmax = 4.8 6.15 = 30 mm. Alternative 3 equivalent linear resistance: For large deformations the stiffness is fairly linear. Assume that the average stiffness is linear and equal to 65 % of the elastic stiffness, i.e. k = 0.65 123 = 80 MPa/m. In this case the rel can be set arbitrarily, but it should be ensured that the response is such that ymax/yel < 1.0, and it is practical to select a given rel/fmax ratio for which a response curve is provided. Hence assume rel/fmax = 1.5, which gives rel = 47.3 mm and then it follows r. The

Consequently; the stiffener is not strong enough to resist the explosion pressure without rupture (see discussion in Section 9.3.2 as concerns rupture of stiffener). It is a fairly common experience that stiffeners are more likely to be critical with respect to explosion loads than the plating between stiffeners.

9.3.1 Plates. Generation of elasticplastic resistance curve is illustrated for a plate with the following particulars: Length, l = 2 m, width, s = 0.5 m, thickness, t = 10 mm, yield stress f y = 355 MPa. It is assumed that the plate is a part of a continuous plate field. Large deformations are expected so that the plate will yield along the boundaries. Then clamped boundaries are assumed. The rigid plastic curve is given by Equation (8.3). The collapse resistance in bending is rc = 0.76 MPa. The resistance curve for fully fixed boundaries are indicated by the line Plate c = inf in Figure 9-6. Below, the curve will be adjusted for the effect of in-plane flexibility using the procedure described in Section 6.8.2. First, the resistance of a plate-strip is calculated, using information given in Section 6.9.2 with = 2 (rectangular crosssection). Clamped boundaries with c1 = 2 are assumed also for

wc =

WP

A

2 t 2 1 4 t = t 1 2

Resistance [Mpa]

the plate strip. The collapse resistance in bending for the plate strip is rc = 0.57 MPa. The characteristic beam height is.

5 4 3 2 1 0 0 10 20 30 40 50 Deformation [mm]

Plate c = inf Plate c = 1.0 Plate c = 0.3 Strip c = inf Strip c = 1.0 Strip c = 0.3

The resistance curve for the plate strip is shown in Figure 9.6 for fully fixed boundaries c = , and for two values of the nondimensional spring stiffness, c = 1.0 and c = 0.3. It is observed that the difference between the plate strip and the plate solution is small for the present fairly large aspect ration, notably when the membrane effect predominates On the assumption that the plate experiences the same relative reduction of the resistance due to axial flexibility as does the plate strip, resistance curves for the plate with non-dimensional spring stiffness, c = 1.0, and c = 0.3 can be generated as shown in Figure 9-6. The next step is to assess the flexibility factor c: If the flexibility of the adjacent structure is neglected, accounting only for the 2nd term in Equation (6.11), there is obtained

k= 2 EA 2 Et 1 = = 8400 MN / m l s

Plate c = inf Plate c = 1.0 Plate c = 0.3

Uniform stress field applied along boundary of removed plate

Inward displacement

Figure 9-7 Elastic-plastic resistance for a plate with various degrees of axial flexibility.

Figure 9-5 Approximate determination of flexibility by means of membrane analysis

In order to assess the influence of the flexibility of the adjacent structure, a membrane analysis is performed with the plate removed, see Figure 9-5. A constant stress of 100 MPa is applied perpendicular the boundaries. The maximum deformation obtained, at the mid-point of the long edges, is 0.25 mm. This yields an equivalent stiffness of knode = 1000.0101/ 0.2510-3 = 4000 MN/m. When both effects are accounted for, the resulting stiffness becomes k = (1/8400 +1/4000)-1 = 2710 MN/m and c = 0.31. Hence, the plate resistance may be assessed reasonably well by means of the curves for either c = 1.0 or c = 0.3. Finally, the linear elastic solution up to the collapse resistance in bending, rc, is added to the rigid-plastic solution. Using the information given in Section 6.9.1, = 400, and k1 = 123 MPa/ m. The deformation corresponding to r = rc is wel = 6.15 mm. The resulting resistance curves are shown in Figure 9.7.

The plate considered in Section 9.3.1 is stiffened with HP 180x 8 stiffeners with yield stress fy = 355 MPa. The girder spacing is 2.0 m. It is assumed that the stiffener is continuous, so that yield hinges can form at the connections to the girder, hence c1 = 2. The area of the stiffener As= 1.8810-2 m2 and the distance to the centroid is zg = 0.109 m. From Figure 8-18 it is found that the plate flange is approximately 80% for a uniformly distributed load when l /s = 0.62.0/0.5 = 2.4. The effective area of the plate flange is 0.8 s t = 410-3 m2 > As. Hence, it may be assumed that the plastic neutral axis for the effective section lies at the stiffener web toe. This yields the plastic section modulus WP = As zg = 2.0510-3 m3 and collapse resistance in bending

R0 = 8c1 f yWP l = 0.58 MN

The characteristic beam height is. WP w c = ----------- = z g = 1.2 0.109 = 0.13 m A The moment of inertia for stiffener with effective plate flange is I = 2.28 10-5 m4. The initial elastic stiffness is taken from Table 6-2: 384EI k= = 230 MN/m L3 This yields a lateral elastic deformation of wel = 2.5 mm for R = R0. The resistance curve for the stiffener with associated plate flange is shown in Figure 9.8 for various degrees of axial flexibility (Note elastic part not included!).

For uniformly loaded, clamped beams there will be an elastoplastic bending phase between the occurrence of first plastic hinge and final formation of final collapse mechanism. To account for this effect, the initial stiffness may be modified on the basis of equal area principle. The equivalent elastic stiffness is obtained from Table 6-2 with m1 = 1: k= 307EI = 184 MN/m L3

and wel = 3.2 mm for R = R0. It is noticed that the stiffener must undergo a substantial plastic deformation before membrane strengthening becomes significant according to the present model. Whether this is achievable depends on the ductility of the stiffener, refer Section 9.4.2. Recent investigations indicate that the model adopted for stiffened plate is considerably conservative, which may warrant a more accurate nonlinear finite element analysis if the stiffener response becomes critical.

2.0 c = inf 1.5 R [MN] c = 1.0 c = 0.5 c = 0.2 c = 0.1 = 1.2

The plastic bending resistance is 8M Pm Rel* = = 5.95 MN L and wel* = 21.8 mm. However, the functional loads amount to 1.8 + 0.29 = 2.09 MN (including steel weight), so 21.82.09/ 5.95 = 7.6 mm is already utilised and only Rel = 5.95-2.09 = 3.86 MN and wel = 14.1 mm is available in the equivalent elastic range. The limiting deformation for rupture calculated in 9.4.3 is wmax = 95 mm, yielding ductility ratio = w/max / wel = 95/14.1 = 6.7. When calculating the load-mass factor the change in transformation factor from the elastic to plastic regime may be accounted for, see Ch.8, Comm. 6.4. The factor for distributed mass and concentrated mass is klmaverage,u = (0.55 + (6.7 1) 0.56) / 6.7 = 0.56 and average,c klm = (0.83 + (6.7 1) 1.0) / 6.7 = 0.975, respectively. The eigenperiod becomes 0.56 2.9 10 + 0.975 1.8 10 T = 2p ---------------------------------------------------------------------------------- = 0.166s 6 274 10 and hence td/T= 0.33/0.166 ~ 2. From Figure 6-3 there is read Rel/Fmax = 0.7 for coordinates (2,6.7). Hence, the girder can resist a dynamic load of Fmax = 3.86/0.7 = 5.5 MN, corresponding to a peak pressure of fmax= 0.23 MPa. Example girder: The neutral axis for the girder studied in Section 9.3.3 is located 0.315 m from the plate flange. This yields a characteristic dimension dc = 2 (1.5 0.315) = 2.37 m. The critical location is at the clamped side, whereby =1/3. Clamped end yields c1 = 2 for the fracture check. With H = 0.0034 and c = 0, there is obtained w/dc = 0.069 and w = 0.095 m.

4 5

1.0

0.5

Figure 9-8 Resistance curve for stiffener with associated plate flange.

9.4.1 Plating Rupture of the plating for the example considered in Section 9.2.2 may be estimated by means of the procedure given in Section 3.10.4, using the plate strip analogy. The characteristic dimension is, dc = t = 10 mm. For steel grade S 355 a strain hardening coefficient of H = 0.0034 is used, refer Table 3-4. = 0.5, c1 = 2 (clamped ends) and l /dc = 0.5 s/t = 25. This yields the following values for the critical deformation, wcrit, depending on the spring stiffness c, see Table 9.1 (Note: the elastic deformation rel = 6.15 mm is added to the values obtained). By inspection of Figure 9-7 it is noticed that the fully plastic membrane state according to this procedure is attained in all cases but c = 0.

Table 9-1 Ductility limit as a function of the spring stiffness c 1.0 0.3 0 wcrit [mm] 35 51 59 76

What is the maximum pressure a steel girder can resist prior to rupture, when the explosion load is triangular, with equal rise and decay time, and the duration is 0.33 s? The girder has the following dimensions: Length L = 12 m, web height, hw = 1.5 m, web thickness, tw = 13 mm, top flange breadth, btop = 0.45 m, top flange thickness ttop = 19 mm. The girder spacing is 2 m and the plate thickness is 10 mm. For simplicity it is assumed that the plate flange is fully effective. The girder has a distributed load of intensity 10 kN/m2 and mounted equipment with mass 1.8105 kg. The equipment load acts equally at two points located L/3 from member ends. The girder is simply supported at one end and clamped at the other end. At the clamped end fully plastic bending moment of the girder can be assumed. There is no axial restraint. Yield stress fy = 355 MPa, acceleration of gravity g = 10 m/s2, density of steel 7.86103 kg/m3. The following is obtained for the girder: Moment of inertia I = 1.8410-2 m4, elastic section modulus, W = 1.9610-2 m3, plastic section modulus, WP = 2.5110-2 m3, total cross-sectional area 0.048 m2. The total distributed mass, including mass of girder is 0.2910-5 kg, so the concentrated mass predominates. Hence, transformation factors for two concentrated loads in Table 6.2 are used. The equivalent stiffness in the elasto-plastic range (m3 = 1) is. 122EI k= = 274 MN/m L3

9.4.2 Stiffener: Rupture is calculated for the stiffened plate considered in section 9.2.3 using the procedure given in Section 3.10.4. The steel grade is S 355 with a strain hardening coefficient of H = 0.0034, refer Table 3-3. Clamped conditions are assumed, i.e. c1 = 2. The shape factor (somewhat arbitrarily) set to 1.5. The characteristic dimension of the stiffened plate is dc = 2hw = 0.36 m. This yields /dc = 5.56, only. This critical deformation becomes wcrit = 0.1dc = 36 mm, almost independent of the spring stiffness c (Note: ductility ratio is = 36/2.2 = 16). This fairly small value is due to the low /dc ratio for the stiffener. The stiffener is far from entering the membrane stiffening phase, so that any discussion of the possibility for membrane forces to develop is irrelevant.

If the stiffener is free against rotation and/or has a longer span membrane effects may become important prior to rupture. Observe that rupture is calculated for the location subjected to the largest strains, i.e. at the stiffener top flange. Rupture in the top flange is not necessarily critical with respect to intactness to explosion loads, because the plate side experiences far less strains. It is likely that the plate will remain intact beyond the deformation limit corresponding to rupture in the top flange. A significant part of the contribution to resistance from the stiffener is lost, but the plating between girders may have a significant residual resistance after failure of stiffeners provided that the plate does not disintegrate. It is, however, difficult to provide validated, closed form solution for this situation. A stiffener subjected to pressure on the plate side may trip about the weld toe at mid span. In this case the assumptions used in the strain calculation model are no longer valid.

9.4.3 Girder:

Plate thickness: t = 14 mm Stiffener dimension: HP240x10, simulated as an L-profile with dimension L240x39x10x29 Stiffener spacing: s = 800 mm l = 3200 mm Stiffener length: Girder dimension: T-girder with dimension: 870x300x10 x20 Girder length: L = 12000 mm The material properties are as follow: Yield strength: Strain rate factor: Effective yield strength: Modulus of elasticity: Material density: Poissons ratio: Max. plastic strain: fy = 420 MPa = 1.0 fy = fy = 420 MPa E = 2.1105 MPa = 7850 kg/m3 = 0.3 1.0% (maximum allowable, correspond to cross section class 3 or 4, see sub-section 9.5.2)

The neutral axis for the girder studied in Ch.8, Comm. 6.10 is located 0.315 m from the plate flange. This yields a characteristic dimension dc = 2 (1.5 0.315) = 2.37 m. The critical location at the clamped side, whereby =1/3. Clamped end yields c1 = 2 for the fracture check. With H = 0.0034 and c = 0, there is obtained w/dc = 0.069 and w = 0.095 m.

Permanent loads and live loads are as follow: Permanent loads: Live loads: Explosion pulse period: pP = 10.0 kN/m2 pL = 5.0 kN/m2 td = 0.15 sec (triangular load with a rise time = 0.50td)

9.5.1 Geometry, material and loads

The geometry of the structure is outlined in Figure 9-4. The main dimensions are:

Stiffener:

10 800 (typ.) Bulkhead 29 39 240

t = 14

Girder:

10

Bulkhead

20

3200(typ.)

300

9.5.2 Cross sectional of properties for the girder Effective plate flange according to DNV Classification Note 30.1 (July 1995), sub-section 3.4.3 and 3.5.4:

=

1 .8

s t

fy E

Cx =

0 .8

2 2 1 1 3200 l l e = s C x + 0.1 1 1 = 800 0 . 58 + 0 . 1 1 1 + + = 784.6 mm 2 2 800 2.56 s Determination of cross section class, Ref. NS3472:2001, Sec- tion 12.1:

Web:

(h wg / t wg ) 235 / f y

( t fg

) (

= 113.6

0.5 (b fg t wg ) 235 / f y

Bottom Flange:

(

Plate Flange:

0.5 (l e t wg ) t 235 / f y

Gross sectional properties: Effective area of plate flange: Ap = let = 303.214 = 4245.1 mm2 Area of girder flange: Af = bfgtfg = 30020 = 6000.0 mm2 Total area of girder web: Aw = hwgtwg = 85010 = 8500.0 mm2 Total area (gross section): AG = Ap+Af+Aw = 4245.1+6000+8500 = 18745.1 mm2 Distance to neutral axis (from bottom of girder flange): t fg h wg t Af + Aw + h wg + t fg 2 + t fg + A p 2 2 z0 = = 403.6mm AG

Web height in tension: ht = z0-tfg = 403.6-20.0 = 383.6mm Web height in compression: hc = hwg-ht = 850.0-383.6 = 466.4mm

Elastic moment of inertia (gross section): t fg 2 Effective girder web according to NS3472: Elastic buckling stress 1 2 2 IG = Af t 2 + Af fg + Aw h wg + A p t 12

h + Aw wg + t fg 2

2 t 2 9 4 + Ap + h wg + t fg AG z 0 = 2.407 10 mm 2

f e = 23.9

t wg 2 E 2 12 (1 ) hwg

Web slenderness:

p =

fy fe

hc hce = hc p

if 1 1 5 p if

hce

ht

ht

Effective girder cross section properties Reduction in web height: h = hc -hce = 466.4 430.8 = 35.6 mm Effective cross section area: Ae = AG -h twg = 18745.1 35.610.0 = 18389.1 mm2 Distance to neutral axis from bottom of girder flange: h + hce 35.6 + 430.8 AG z 0 h t wg + ht + t fg 18745.1 403.6 35.6 10 + 383.6 + 20 2 2 z 0e = = = 399.1mm Ae 18389.1

I Ge = I G h 1 h 3 t wg h t wg t fg + ht + c z 0 e 12 2

2

I Ge = 2.407 10 9

W eu = I Ge 2.387 10 9 = = 5.982 10 6 mm 3 z 0e 399.1

W e = min(W eo , W eu ) = 4.923 10 6 mm 3

Plastic section modulus:

Web areas:

Aw1 =

e1 h ce

Aw 2

e3 >

Aw1

e1 =

6000 + 2154.0 + 5990.0 4245.1 = = 494.9mm 2 10 = 6000 2145.0 + 5990.0 4245.1 = 279.5mm 2 10

h ce

e3

e3 =

Aw2 ht

hc 2 + ht if e2 = 2 e3 if

hc 2 + ht 2 h e3 c 2 + ht 2

e 2 = 279.5mm

2

= 8.719 10 6 mm 3

e t W p 2 = A p + e1 + A f (t fg + h wg e1 ) + 1 t wg 2 2

hce + ht e 2 2 + t wg = 1.812 10 6 mm 3 2

2

h e W p 32 = Aw1 ce + h + e 2 + 2 t wg 4 2

W p 3 = W p1 + W p 2 = 6.070 10 6 mm 3

Plastic section modulus: W p1 if A p > Aw1 + Aw 2 + A f W p = W p 2 if A p + Aw1 > Aw 2 + A f = 6.070 10 6 mm 3 W p 3 if A p + Aw1 < Aw 2 + A f Ratio between plastic and elastic section modulus: Wp = 1.23 We

9.5.3 Mass

Mass from plate:

Mass from stiffener, see figure:

kg m

tws = 10 hws = 240-29 = 211

As = hws t ws + b fs t fs = 211 10 + 39 29 = 3241mm 2 ws = As 3200 l 3241 kg = 7850 = 101.8 800 s 10 6 m kg 18745.1 7850 = 147.1 6 m 10

tfs = 29

w g = AG =

bfs = 39

Mass from permanent loads and possible live loads (to be evalp kg 10 10 3 w PL = P l = 3.200 = 3263.1 9.807 g m Total mass:

9.5.4 Natural period Linear Stiffness, Ref. Table 6-2 in Section 6.10:

kl = 384 E I Ge L3 =

kg m

Natural period assuming uniformly distributed mass (Klm,u is taken from Table 6-2):

T = 2 K lm ,u M u kl = 2 0.77 w L 0.77 3863.7 12.0 = 2 = 0.113 sec kl 1.114 10 8

td 0.15 = = 1.33 T 0.113

9.5.5 Ductility ratio The maximum lateral deformation prior to buckling can be calculated according to equation 3.19 in sub-section 3.10.2: 2 wp 14 c f f y L 1 = 1 1 dc d 2c f c1 3 c

Web (bending):

= 0.8

h wg / t wg

235 / f y

= 0 .8

= 90.9

where; dc is characteristic dimension for local buckling, i.e. 2(t+hce+dh+e3) = 2(14+430.8+35.6+279.5)= 1089mm c1 is 2 for clamped beams L is the smaller the distance from load to adjacent joint (0.5). Here set to 0.5L, i.e. 6000

Based on these input parameters, the maximum plastic deformation is calculated to: 2 1089 14 0.994 420 6000 wp = = 33.37 mm 1 1 2 0.994 1089 2 90.9 3 The maximum elastic deformation is found from:

c 106624 cf = 1 + c = 1 + 106624 = 0.994 ,and c is non-dimensional spring stiffness, ref Section 3.7; c= 4 c1 k wc2 4 2 7.873 10 9 406.12 = = 106624 f y Ae l 420 18389.1 12000

1 1 k node 1 + 2 E AG = 1 = 7.873 10 9 1 1 + 1 10 20 2 2.1 10 5 18745.1

we =

f y W e L2

32 E I Ge

p

M e = f y We = p L2 12

k =

we = 1 p L4 384 E I

knode is axial stiffness of the node with the considered member removed, here assumed infinitely. 1.2 W p 1.2 6.070 10 6 = = 396.1 wc = 18389.1 Ae Calculation of cross sectional slenderness factor, ref. Section 3.10, i.e. the maximum of the following: Plate flange: le / t 303.2 / 14 =3 =3 = 86.9 235 / f y 235 / 420 Bottom flange:

we =

f y We L2 1 p L2 12 L2 1 L2 = Me = 384 12 EI 32 32 E I EI

Ductility ratio:

wp we

=3

b fg / t fg

235 / f y

=3

= 60.2

9.5.6 Maximum blast pressure capacity From Figure 9-11, the dynamic load factor is found:

DLF ( ) = Rm = 0.99 Fl

With reference to Figure 9-11, k3 was set to 0, which ensures conservative results.

100

Rel/Fmax=0.05

=0.1

= 0.3

= 0.5

= 0.6 = 0.7

Rel/Fmax= 0.8

10

ymax/yel

= 0.9

= 1.80

= 1.0 = 1.1 = 1.2 = 1.5

F

k3 = 0 k3 = 0.1k1 k3 = 0.2k1 k3 = 0.5k1 0.1 0.1 1

R Rel

k3 = 0.5k1 =0.2k1

=0.1k1

Fmax td

k1 0.50td

yel

y

10

td/T

td/T = 1.33

Figure 9-11 Dynamic Response of a SDOF system due to a triangular pulse load profile (rise time = 0.50td)

Maximum resistance for a fixed supported beam, see Figure 9-12: 16 M p 16 W p f y 16 6.070 10 6 420 Rm = = = = 3399.2 10 3 N = 3399.2kN L L 12000

Elastic p Plastic pP

MP = pPL2/16 MP = pPL2/16

Resistance utilised in order to take the permanent and live load: R 0 = w g g L + ( p p + p l ) l L = 147.1 9.807 12 + (10 + 5) 10 3 3.2 12 = 593.3 10 3 N = 593.3kN The maximum blast pressure capacity is obtained from the following two equations:

Fl = Rm R0 DLF ( ) and Fl = pmax l L

pmax = Rm R0 1 (3399.2 593.3) 103 1 = = 0.074 MPa = 0.74bar DLF ( ) l L 0.99 3200 12000

Note that the maximum resistance (Rm) given above does not include a capacity check with respect to shear. The shear capacity can be determined from sub-section 12.4.4 in NS3472-2001. DET NORSKE VERITAS

- Theory of Local BucklingUploaded byTricia Lee Cairns
- DNV-RP-C203Uploaded byNoralia Amanina Osman
- oto01063[1]Uploaded byurmisharjah
- HAZID Guideword ChecklistUploaded byqhseconsult
- Structural ReviewerUploaded byJireh Grace
- Dnv Dal Spec.Uploaded bygambin888
- Offshore ManualUploaded byengineeringyusuf
- biomechanics of the mandible.pdfUploaded byRhonda Bush
- AccidentalUploaded byMUTHUKKUMARAM
- Stress Strain& Stress Concentration FactorUploaded by2gangadhara
- Tcm StrengthUploaded byfishatsyon
- 2001Artigo_Wang-Hsu_Nonlinear Finite Element Analysis of Concrete Structures Using New Constitutive ModelsUploaded byMarcela Bruna
- EG-120Uploaded byMikey Haken
- Circular Concrete-Filled Tubular Columns- State of the Art Oriented to the Vulnerability AssessmentUploaded byDarsHan MoHan
- Wp1Uploaded byRoberval Pimenta
- Deformations of Reinforced Concrete Members at Yielding and UltimateUploaded byAnonymous b5e3wqzLEJ
- [Dr._Mirvat_Bulbul]_ENCE335_Fundamentals_of_Reinfo(bookzz.org).pdfUploaded byCY Lee
- Civil_Reinforced Concrete StructuresUploaded byPreetha Sreekumar
- deshpande.pdfUploaded byOmkarKocharekar
- 6cb5da3513bd26085ee3fad631ebb37a-original.pdfUploaded byPriyaKumar
- EM302 Stress Analysis and Design May 2012Uploaded by陈韦权
- REZI_2corectatUploaded byBza ZaBug
- DNVRP-C204Uploaded byjuliodevelas
- KA20903 - Lecture 1 Introduction to Mechanics of SolidsUploaded byKit Meng Lim
- JCSR 57 (2001) - 491 - APT,CGS - Strength of Compressed Rectangular Plates Subjected to Lateral Pressure.pdfUploaded byGanesh.Mahendra
- 02mainUploaded byiSoK11
- SCIENCEUploaded byJames Jamsmith
- CIE4115 - Exam 1 November 2013 v3-FinalUploaded byTim Syki
- A Technical Seminar OnUploaded byanon_843262993
- DT for Sandwich StrutUploaded bySanjib Rai

- Collision Resistance of Ship Shaped Structure to Side ImpactUploaded byMuhammad Habib
- Small and Medium Scale Lng Terminals WartsilaUploaded byMuhammad Habib
- Geo5 Engineering ManualsUploaded byMuhammad Habib
- vetus battery chargerUploaded byMuhammad Habib
- BILGE SYSTEM CALCULATIONUploaded byMuhammad Habib
- Ship bed catalogue, marine bed catalogueUploaded byMuhammad Habib
- 2719-5949-1-SMUploaded bybetahita_80174

- Formal Lab - Exp.4 - Chemical Kinetics - CHM 1311 by Simon HagosUploaded bySimon Hagos
- Mechanical and Swelling Properties of Hydrogel.pdfUploaded byThinh Nguyen Tien
- lithosphere hazard compressedUploaded byapi-323848578
- Metrology and Measurements - Lecture Notes, Study Material and Important Questions, AnswersUploaded byM.V. TV
- Mbma SeismicUploaded byAlexander Tavarez Santana
- Chap 003Uploaded byNguyen Thi Thanh Nha
- Comparative Analysis of Heat Sink Pressure Drop Using Different MethodologiesUploaded byzahid_497
- Luigia Binda-Learning from failure_ long-term behaviour of heavy masonry structures -WIT Press (2008).pdfUploaded byFerenczi Z. Sámuel
- 45662597562Obj CivilEngineering 2006Paper IUploaded by94738183
- Edm Test PaperUploaded byazeemdcet
- MMT-A-2008-39A-3Uploaded byrajkumar
- Sensing Mechanism of Hydrogen Sensors Based on Palladium-loaded Tungsten Oxide (Pd WO3)(2012)Uploaded byAlex Faudoa
- Thermal Physics Lesson PlanUploaded byEdwin Tan Pei Ming
- 2009-_Manufacturing_of_membrane_electrode_assemblies_for_fuel_cells.pdfUploaded byNardos Henok
- Investigating Factors Affecting Level of Capillary Action COURSEWORKUploaded byAlexander Chen
- PoroElasticity2013.pdfUploaded byyehuoy
- Seminar ReportUploaded byRamaswamy Koodalloor Parasuraman
- Failure Analysis in Material ScienceUploaded bymaylynXiX
- Bs 2762 Elastometric Bearing Seimic IsolatorsUploaded byFrancis Zigi
- G484 - The Newtonian World Revision CardsUploaded byThomas Roe
- Moment of InertiaUploaded byssjuned
- End Point of Wet Granulation by Measuring Powder Energies and Thermal PropertiesUploaded byckumar2
- SAES J 700 Control ValvesUploaded bygoldengineer14
- 12 Physics Chemistry EMUploaded byHameed Bin Ahmad
- Microscope Leica DmlmUploaded byyaku1618
- SOLUBILITY ENHANCEMENT TECHNIQUES: REVIEWUploaded byIjupbs Ijupbs
- 625 12 Chemistry Minimum Study MaterialUploaded bypriya
- ME2134-1 Lab Manual (Stability of Floating Body)Uploaded bydasdas
- Colombian Olympiad PhysicsUploaded bysanits591
- ISO 11146Uploaded byDanny