Structural Fire Engineering: Introduction with max DOI.pdf · Structural fire safety is therefore a key consideration in the design of buildings and is attracting worldwide attention

Peer-reviewed by international ex-perts and accepted for publication by SEI Editorial Board

Paper received April 17 2012Paper accepted June 2 2012

Structural Engineering International 42012 Scientific Paper 449

The Role of Connections in the Response of Steel Frames to FireIan Burgess Prof John Buick Davison Senior Lect Gang Dong Shan-Shan Huang Lect Department of Civil and Struc-

tural Engineering University of Sheffield Sheffield South Yorkshire S1 3JD UK Contact ianburgesssheffieldacuk

DOI 102749101686612X13363929517811

Abstract

Connections are critical structural elements of building frames and in a fire are subject to forces very different from those at the ambient temperature for which they are designed The fracture of a connection can cause the collapse of the con-nected beam which may lead to a progressive collapse sequence affecting the entire building This paper overviews the sequence of research on connection behaviour in a fire at the University of Sheffield Early work focused on studying con-nections in terms of their momentndashrotation behaviour alone Concurrent full-scale building fire tests led to the realization that the tying capacity of connections is of prime importance for maintaining the structural stability in a fire For whole-structure numerical modelling in performance-based fire engineering design the development of the component-based approach which was initially introduced for ambient temperature connection design is an appropriate way to rationalize and model connection behaviour under these complex loadings The effect of high co-existent rotation on the tying capac-ity of connections has been studied in furnace tests at various temperatures which have provided data to assist in the characterization of the component-based model A general component-based connection element into which appropriate component models can be inserted has been developed so that full connection performance including fracture of compo-nents can be integrated into global non-linear structural fire analysis This will allow buildings to be modelled for a range of fire scenarios so that they can be designed to avoid progressive collapse in a fire

Keywords steel structures fire connections robustness catenary action component method global analysis

Fires in buildings may have enormous consequences on life safety and economy Structural fire safety is therefore a key consideration in the design of buildings and is attracting worldwide attention Significant advances in research have increased the knowledge on the structural behaviour in fire and fire safety engineering has become a highly regarded dis-cipline integrating all aspects of fire safety into the design of buildings

To exploit the innovations and advances gained recently Structural Engineering International launched towards the end of 2011 a call for papers on the topic of Structural Fire Engineering and received a great response from around the world In this issue 10 Scientific Papers and one Technical Report are presented The first papers deal with steel and composite structures with focus on the material and struc-tural behaviour of connections in fire which is important for the overall stability and safety of buildings While some areas like the behaviour of steel structures in fire are becom-ing well understood others like for example the fire behav-iour of concrete structures need further studiesmdashtrends that are reflected by the contribution to this issue with two

scientific papers one of them presenting the state of art of concrete structures that incorporate FRP Because of the combustibility of wood most building codes have strictly limited the use of timber as a building material in the past in particular by limiting the number of storeys The better knowledge in the area of fire design of timber structures from research projects currently allows the safe use of tim-ber and a wider field of application of timber for build-ings An overview and new developments in this area are presented in two scientific papers The last two papers of this special issue are related to case studies of buildings where fire safety engineering and fire performance - based design concepts were successfully applied allowing an optimisation of the structure

The papers will provide a good overview and update on tech-nical and scientific activities achievements and trends in the field of fire safety engineering

Andrea Frangi SEI Editorial Board Chair of WC2

Switzerland and

Markus Knobloch Member of WC2 Switzerland

Structural Fire Engineering Introduction

450 Scientific Paper Structural Engineering International 42012

joints that are assumed to be pinned at ambient temperature can provide sig-nificant levels of strength and stiffness at elevated temperatures especially as their temperatures remain lower than those of the connected beams They can therefore reduce significantly the deflections of these beams at any given stage of a fire which increases their fire resistance in conventional5 fur-nace testing It is clear that the internal moments in beams can be redistrib-uted to adjacent cooler members as was illustrated in the Broadgate fire report6

Experimental investigations con-ducted on the performance of steel joints at elevated temperatures are relatively recent and rather limited in number mainly because of the high cost of fire tests and the limitations on the size of the furnace used The joint tests conducted have primarily focused on establishing momentndashrotation rela-tionships of isolated joints However even on these terms an experimental fire test is incapable of extrapolation to details other than those used in the actual test or to other fire timendashtem-perature relationships Results from fire tests of i solated joints provide important fundamental data on the behaviour of joints but do not truly reflect the actual behaviour of joints in buildings in the event of a fire due to the absence of structural continu-ity Thus in recent years the numeri-cal simulation of joints and structural frames subjected to fire conditions has been the basis of extensive research7 This paper attempts to trace the devel-opment of research work that initially concentrated on investigating the rota-tional behaviour of beam-to-column joints at elevated temperatures This experimental work on isolated joints has le d to the development of various analytical approaches to predict the r otational behaviour of both bare-steel and composite joints and the effect of structural continuity on the perfor-mance of beams supported by realistic joints

Rotational Behaviour of I solated Joints in a Fire

The first experimental fire tests on joints were conducted by Centre Technique Industriel des Constructions Metalliques (CTICM)8 on six types ranging from ldquoflexiblerdquo to ldquorigidrdquo The primary purpose of these tests was to investigate the performance of high-strength bolts at elevated temperatures

Introduction

The mechanical properties of steel structures de grade rapidly in a fire because of the reduction in both the stiffness and strength of the material at high temperatures Using applied fire protection (covering the exposed steel with a prescribed thickness of an insulating material) remains the most common way of satisfying the fire resistance requirement for a steel-framed structure However it is more rational to consider fire as an addi-tional load case1 just as designers routinely consider wind or earthquake hazards and to design the structure with appropriate protection to be acceptable at the fire limit state rather than designing the structure for other limit states and then applying retro-spective fire protection Following the recognition of this logic interest has grown in understanding the behaviour of structural elements in a fire both in isolation and as part of the whole building Experimental and theoretical research has made considerable prog-ress towards the goal of considering the fire limit in the normal analytical design process

All structural members exposed to fire heat up but with different rates of tem-perature rise Joints in a steel-framed building tend to heat more slowly than the material within the free span of the beams that they connect because of the additional mass of material (bolts plates angles etc) in a shielded loca-tion (ie usually beneath a composite floor) and a relatively small surface area EN 199 3-1-2 20052 suggests sim-plified joint temperatures of between 62 and 88 of that in the beam lower flange temperature at mid-span Alternatively the temperature distri-bution may be calculated on the basis of the exposed area volume ratio of each joint component (plate angle bolt etc) Franssen3 has suggested that this method may be non-conservative

Traditionally the design of steel-framed structures has assumed that the connection between a beam and a column is either rigid (implying com-plete rotational continuity) or pinned (implying that there is no moment transfer) However actual connection behaviour exhibits characteristics over a wide spectrum between these two limits connections regarded as pinned generally possess some rotational stiffness whilst ldquorigidrdquo connections display some rotational flexibility4 It was realized in the late 1980s that

and no indication of the performance of the joints was reported Two tests were carried out by British Steel9 in 1982 on a ldquorigidrdquo moment-resisting joint which sustained significant deformation during a fire The first tests to assess the structural continu-ity afforded by beam-to-column joints at elevated temperatures were car-ried out by Lawson10 with the aim of developing a design approach for steel beams taking into account the rota-tional restraint provided by end-plate or web angle joints The tests dem-onstrated the strength of these joints in a fire and showed that significant moments (up to two thirds of their ambient temperature design moment capacity) could be sustained in fire conditions Lawson proposed simple rules for designing simply supported beams in a fire taking into account the moment transferred to columns via joints in fire conditions Although the test results provided insufficient data to describe the momentndashrotation char-acteristics of the joints they provided essential information for the early attempts at joint modelling

The first furnace tests intended to develop momentndashrotation relation-ships for flush end-plate joints at elevated temp eratures were con-ducted by Leston-Jones et al 11 Eleven tests were carried out on small-scale specimens including two at ambient temperature for both bare-steel and composite joints The results demon-strated that both the initial stiffness and moment capacity of the joints decreased with increasing tempera-ture with a significant reduction in capacity for temperatures in the range of 500 to 600degC These tests provided useful data for connection model-ling but for a limited range of details employing relatively small section sizes for comparison with earlier joint testing work at ambient temperature by Davison et al12 A further series of elevated temperature joint tests was conducted by Al-Jabri et al13 to study the influence of details such as mem-ber size end-plate type and thic kness and composite slab characteristics on the joint response in a fire For each joint a series of tests was conducted at a constant moment but with increas-ing furnace temperature14

Prediction of Joint Characteristics Using Component-Based Models

The component method was devel-oped following the suggestion1 516 in the 1970s that it represented in a prac-


tical fashion the rotational response of joints at ambient temperature with the objective of facilitating a design of semi-rigidly connected frames It is based on the division of a joint into basic com-ponents of known mechanical prop-erties Each joint component such as the e nd-plate the column flange bolts etc is idealized as a bilinear spring of known stiffness and yield strength The elastic behaviour of the joint may be determined by assembling the stiff-nesses of the individual components to form a global joint rotational stiffness Extensive research work reported by the network COST C117 was devoted to modelling the rotational behaviour of isolate d joints (without consider-ing axial forces) at a mbient tempera-ture using the component method The outcomes of this research were used to develop EN 1993-1-8 2005 18 which includes recommendations for model-ling of joint characteristics using the component method

Rotational response at elevated tem-peratures may logically be predicted by degrading the stiffness and strength of each component in a bolt row according to its temperature allowing the modelling of this behaviour under any temperature distribution through the connection The jointrsquos rotational stiffness and strength are therefore degraded with increasing temperature At elevated temperatures the ini-tial work of studying rotational joint behaviour was conducted by Leston-Jones19 who proposed a simple com-ponent model to model the response of bare-steel and composite flush end-plate joints as well as conducting high-temperature experiments Spyrou later successfully developed models for ten-sion20 and compression behaviour21 of zones of flush end-plat e beam-to-column connections Spyroursquos model for the behaviour of a component is illustrated i n Fig 1 which idealizes the beam and column faces as rigid bars connected by two non-linear springs each of which can act within the ten-sion and compression quadrants shown in Fig 2 The comparison of the resulting model predictions with Leston-Jonesrsquos test data is shown in Fig 3 A similar component model was proposed by da Silva et al22 for bare-steel flush end-plate joints at elevated temperatures which again compared well with experimental results avail-able in the literature Al-Jabri2324 also developed component-based models for bare-steel and composite flexible end-plate joints The comparison of the bare-steel component models with

Fig 1 Spyrou model of joint with axial force and moment

Kt

Kc

ZM

P

Fig 2 Forcendashdisplacement of a joint component (Spyrou20 21)

Ft

Fc

Ft4

Ft3

Ft2

Ft1

c1

t1 t2 t3 t4

c2c3c4c t

Fc1

Fc2

Fc3

Fc4

Kt1

Kc1

Kc2

Kc3

Kc4

Kt2

Kt3

Kt4

Fig 3 Comparison of Leston-Jones tests19 with the Spyrou model

900

800

700

600

500

400

300

200

100

0 10 20 30 40 50Rotation (millirads)

Bea

m fl

ange

tem

pera

ture

(degC

)

60 70 80 90 100 110 120 130

5

1015 20

25

Connection

Model

Test

Moment (kN m)

existing test data was especially good in the elastic zone and the predicted deg-radation of joint stiffness and capacity compared well with the experimental results For composite joints the pre-dicted and measured responses agreed well at ambient temperature and were encouraging at elevated temperatures

although more extensive test data was (and still is) required

Performance of Joints in Frames in a Fire

In the ambient temperature context joints are usually assumed to resist


have sufficient flexibility to allow rota-tion Important observations from the Cardington fire tests4243 included the following

bull Despite the partial (one-sided) frac-ture of the fl exible end-plates which probably occurred during cooling the connected beams showed no sign of collapse under very high defl ections (Fig 4)

bull The temperature of the bottom fl ange of the beam was considerably higher (as much as 200degC) during heating than the mean temperature of the joint The temperature of the bottom bolt row was higher than that of the top bolt row and the end-plate was hotter than the bolts at the same level

bull The joints were subjected to high tensile forces In the fl exible end-plate joints the plates had frac-tured down one side adjac ent to the weld while the other side remained intact as shown in Fig 5(a) In the fi n-plates (beam-to-beam joints) the bo lts often sheared (Fig 5(b)) These fractures occurred as a result of the high tensile force s developed during the cooling of the connected beam However such behaviour was not observed in isolated member fi re tests44 The behaviour of joints

on restrained sub-frames employing a range of connection types

Lessons Learnt from the Cardington Full-Scale Frame Tests

It has often been observed that com-plete steel-framed bui ldings tend to behave better in accidental fires than would their individual members tested in isolation because of the interac-tions between structural members In order to observe the behaviour of a real building under natural fire con-ditions and to collect data that would allow verification of analytical mod-els proposed for analysing structures during a fire the Building Research Establishment conducted a series of fire tests on a full-scale composite building structure at Cardington UK constructed in 199437 A full descrip-tion of the Cardington full-scale frame and its associated tests is presented in detail elsewhere38ndash41

Two types of joints were used in the Cardington frame flexible ( partial depth) end-plates and fin-plates Flexi-ble end-plates were used for beam-to-column joints and fin-plates for beam-to-beam joints These joints are usu-ally considered as pin joints they are assumed to transfer shear forces and to

vertical shear forces but may have rotational properties ranging from frictionless rotation to full fixed-end moment transfer The major research efforts in recent times have been aimed at establishing their momentndashrotation response without considering the concurrent thrust parallel to the axis of the beam In some steel struc-tures such as pitched-roof portals and unbraced continuous multi-storey frames the magnitude of the axial forces generated within the beams is significant2 526 and this affects the ambient temperature performance of joints The effect of axial forces is insufficiently addressed in EN 1993 -1-8 which only suggests an em pirically based limitation on the allowable axial force of 10 of the beamrsquos axial plas-tic resistance below which the effect of the axial force can be ignored

Several studies have been carried out on the effect of axial force on initial rotational stiffness and to establish bending momentndashaxial force (MndashN) interaction curves for different joint configurations using the compo-nent method27ndash29 Wald and S varc30 and Luciano de Lima et al3132 stud-ied experimentally the behaviour of beam-to-column joints in the presence of an axial force In the former study two tests were performedmdashon beam-to-beam and beam-to-column jointsmdashwhilst the latter examined 15 joint configurations (8 flush end-plate and 7 extended end-plate joi nts) These stud-ies confirmed that the presence of an axial force can significantly affect the jointrsquos structural behaviour

The effect of axial thrust on the behav-iour of joints is more critical when steel structures are subjected to fire Beams expand significantly due to thermal expansion and contract when the structure cools causing high axial thermal stresses if these movements are resisted The normal forces on connections at the beam ends can sig-nificantly affect the behaviour of these connections Experimental and ana-lytical studies33 34 performed on a sub-frame assembly concentrated on the effect of axial restraint on the behav-iour of steel beams in a fire without giving much attention to the behaviour of joints under moment combined with axial restraint Simotildees da Silva et al35 provide a useful tabulated summary of the work conducted on the influence of end restraint on structural response under fire loading More recently Wang et al36 have reported the results of high- temperature experimental work

Fig 4 Cardington overall frame deformation after a fire

Fig 5 Failure modes o f flexible end-plate and fin-plate joints in Cardington fire tests

(a) (b)


just enough to balance its net tensile capacity against the catenary tension caused by its loading and deflection If the beam is cooled below any tempera-ture the recovery of its thermal expan-sion as the material stiffens generates high tensile tying forces at its ends If the connections or surrounding struc-tures are ductile during this tension phase then the catenary tension will be reduced as will the enhanced tension caused by the cooling

Connections at the ends of heated steel beams are the first link in the load path of these restraint forces and are also potentially the most vulnerable components in the chain very rarely being designed specifically for ductil-ity in tying action In UK practice as in many other countries connections are usually designed as ldquosimplerdquo with the principal role of resisting the verti-cal reactions at beam ends but with a fairly nominal tying (normal tension) strength requirement

Incorporating Joint Behaviour in Finite Element Analysis

Performance-based structural fire engi-neering analysis and design has been used largely to optimize the location and quantity of fire protection materi-als and to some extent it has acquired the image of being used to reduce fire protection costs to developers However for large and complex struc-tures whose design has been optimized to a considerable extent in the context of all the other design limit states there is a much more fundamental reason to use performance-based analysis of the fire limit state It has been shown ear-lier that structural interactions in fire scenarios can be extremely complex and that prescriptive fire protection has (in the case of 7 World Trade) not prevented disproportionate building

It is clear that in most cases the most vulnerable parts of steel and compos-ite buildings in a fire or other haz-ards are the connections between beams and columns These are usually designed to carry forces under ambi-ent temperature loadings that are eas-ily defined and calculated However it has been seen that in fire conditions the response of the connected beams causes a complex variation of forces for which the connections have almost cer-tainly not been designed It is instruc-tive to consider the typical variation50 of ldquotyingrdquo forces (force components perpendicular to the column face) applied by beams to the connections as temperatures rise and fall during the progress of a building fire An example is presented in Fig 7 which shows the tying force component transferr ed through the connections from a beam to the columns at its ends as the beam temperature increases The material properties that influence this varia-tion directly are thermal expansion and strength degradation with tem-perature Heating of a steel downstand beam causes a free thermal expansion which if stiffly restrained (Fig 7(a)) by surrounding structures such as pro-tected columns cooler beams attached concrete slabs or braced bays gener-ates very high compressive forces If the beamrsquos free thermal expansion can be accommodated by a soft ductile surrounding structure then the initial build-up of compression force will be greatly lessened (Fig 7(b))

As temperatures rise further the net compression is progressively reduced by the sagging deflection of the beam and by the loss of material strength and stiffness At very high temperatures nearly all the bending stiffness of the beam has been lost and it hangs essen-tially in catenary tension between its end connections eventually deflecting

during the cooling of the structure clearly needs further investigation

bull Local buckling of the beamrsquos lower fl ange and web occurred during the heating phase (Fig 6) T his buckling was caused by high compression resulting from the restraint on ther-mal expansion provided by the adja-cent cooler structure together with the negative moment caused by the rotational restraint of the joint The results from an analytical study of the fi rst Cardington test45 confi rmed that the response of the structure was mainly dominated by the effects of thermal expansion and that mate-rial degradation and gravity load-ing were of secondary importance Local buckling was found not to be a major concern in isolated member fi re tests42

Forces Imposed on Connections in a Fire

The dramatic collapse of the twin towers of the World Trade Center is an enduring image of progressive col-lapse caused by the effects of fire on buildings that had initially withstood the considerable physical damage caused by aircraft impacts The total collapse later on the same day of a nearby 47-storey building (7 World Trade) which had seemed to have taken relatively minor structural dam-age but had been affected by lengthy internal fires is less well remembered but would in a more normal context have been viewed as a cause for con-siderable concern A series of recent reports46ndash49 have focused attention on the need to design and construct robust structures capable of coping with dif-ferent types of accidental or malicious damage In the case of 7 World Trade in particular it was suggested that the forces applied to connections via the restrained thermal expansion of long-protected steel beams after prolonged exposure to fires caused the local failure that initiated the progressive collapse

Fig 6 Local buckling of beams in the vicinity of a joint

Fig 7 Tying forces in typical beamndashcolumn connections as the beam temperature increases (a) stiff restraint to horizontal movement (b) ductile restraint to horizontal movement

400

(a) (b)

200

0

minus200

minus400

minus600

minus800

0 200 400 600 800 1000 1200

Temperature (degC)

Axi

al fo

rce

(kN

)

400

200

0

minus200

minus400

minus600

minus800

Axi

al fo

rce

(kN

)

Axial force in restrained beam

Steel tensilestrength

Tension

Compression

Heating

Cooling

0 200 400 600 800 1000 1200

Temperature (degC)



Tension

Compression

Heating

Cooling


and Node 2 is the end node of the beam The shear components have not been characterized at this stage and are assumed to be rigid in the vertical shear direction

Tension C omponent

Each tension bolt row includes three components which are connected in series The middle component in each series represents the bolt in tension For a flush end-plate connection the other two compone nts represent the column flange in tension and beam end-plate

connections of different types has been developed and implemented in the global analysis software Vulcan This is a logical development of Blockrsquos model and includes component characteriza-tions that have been developed since 2005 Figure 10 shows a schematic lay-out of the component assembly within the component-based connection ele-ment The assembled element has two external nodes internally it consists of five ldquotensionrdquo component rows and two ldquocompressionrdquo component rows Node 1 coincides with a column node

collapse in a fire The assessment of structural fire resistance in design ought to be based on the use of reli-able computational m odels of whole-structure behaviour subjected to a range of extreme fire scenarios based on agreed risk levels Furthermore this modelling clearly needs to be capable of modelling the connection behaviour and the sequence of failure until local or overall collapse occurs

Since detailed finite element (FE) modelling particularly of connections at this scale would be extremely oner-ous for the designers who have to cre-ate a full-structure model it is clear that analysis based on a more macro-scopic approach will be necessary Such software tools based on specialized beam-column and slab elements that account for temperature profiles and high-temperature behaviour already exist5152 altho ugh their representa-tion of connection characteristics has hitherto been restricted to rotational characteristics Since connections can experience tying forces of significant magnitudes with corresponding defor-mations together with high rotations in a fire the component method offers the possibility of assembling ldquoconnec-tion elementsrdquo that can represent the behaviour of particular connections as part of such analyses The objective is to allow designers to define a con-nection with its engineering informa-tion (type dimensions bolt sizes steel grades etc) which then translates internally into component data and is assembled as a two-node element at the end of each beam

Building on the earlier wo rk by Spyrou Block et al53 further developed a com-ponent model for end-plate connec-tions which includes the end-pl ate in bending the column flange in bendi ng bolts in tension and the column web in compression (see Fig 8) The first three components form the tension zone of the connection and are com-bined as two T-stubs in series A shear spring is included to transfer the verti-cal force at the column face from one node to another this is assumed to be rigid at present although the formula-tion of the element allows the imple-mentation of slip and shear failure of the bolts The model has been vali-dated against the test data by Leston-Jones11 as shown in Fig 9

Assemb ly of Component-Based Elements for Full-Structure Analysis

A component-based connection ele-ment that can be used to represent

Fig 8 Component model developed by Block52 for a connection zone with shear deformation

42 3 1 3

5 Mi Ni ui

Nj uj

Vj wj

jiVi wi

(a) (b)

k1

k2

k3

i Mj j

0 mmw

u

l c2

l c1

lT2

lT3

lT1

Fig 9 Comparison between results from Leston-Jones tests19 and Blockrsquos52 component model

0

100

200

300

400

500

600

700

800

900

0 10 20 30 40 50 60 70 80 90 100

Connection rotation (millirads)

Stee

l tem

pera

ture

(degC

)

Leston-Jones ndash BFEP 10 -10 kN m

Connection element ndash bolt rows as a group

Connection element ndash bolt rows individually

Leston-Jones ndash BFEP 20 ndash 20 kN m



Fig 10 Schematic component-based element assembly

Tension componentCompression component

11 2


respectively at which unloading occurred In Fig 14 the node (FA DA ) is called the intersection point and the intersection of the unloading curve with the zero-force axis is called reference point 1 which represents the permanent deformation caused If the applied force or the componentrsquos displacement is beyond its intersection point its displacement and applied force lie on the loading curve and the permanent displacement increases accordingly On the other hand if they lie below the inters ection point then they are on the unloading curve and the permanent deformation does not change Figure 15 shows how the loading and unloading curves form the ldquoeffectiverdquo FndashD curve representing the componentrsquos behaviour

Unloading with Changing Tempera tures

When a component is heated in a fire its FndashD curve is temperature depen-dent and this temperature changes continuously during the fire The ldquoref-erence pointrdquo concept is introduced to locate the unloading curve The com-ponentrsquos permanent deformation is assumed not to change55 when only the temperature changes When mov-ing to the next temperature step the

plastic deformation compression and residual compression

Compression Spring Row

A compression component is usually represented by three points (Fig 13) A compression component will be ldquoswitched offrdquo under tension when its contribution is zero Point 3 is the ultimate strength beyond which it is assumed that no change of resistance oc curs

Effective ForcendashDisplacement Curve of a Component at Constant T emperature

When a component carries a force it may become inelastic and it acquires irreversible deformation (residual deformation) when its force is reduced to zero In this development the clas-sic Masing rule54 is employed for this ldquomemory effectrdquo The unloading curve is the original loading curve doubled and rotated by 180deg If the initial load-ing curve is represented55 by

D = f(F) (1)

then the unloading curve can be described as

(DA ndash D)

_________ 2 = f (

(FA ndash F) ________

2 ) (2)

where DA and FA (as shown in Fig 14) are the displacement an d force

in tension The forcendash displacement behaviour of each tension component is characterized by a multilinear curve consisting of positive stiffness seg-ments together with a fracture point

The three components in each ten-sion bolt row are combined into one effective spring at each temperature step (Fig 11) After the global analy-sis reaches a converged stable equi-librium the forces in the tension bolt rows are established and the displace-ments of eac h tension component are calculated The related information such as each compone ntrsquos permanent deformation is then updated At each force level the effective springrsquos dis-placement is the total of its compo-nentsrsquo displacements under this force level The typical tension bolt row forcendashdisplacement curve (Fig 12) consists of four par ts tension bolt

Fig 13 Basic model for compression com-ponent forcendashdisplacement behaviour

Displacement

For

ce

Point 2

Point 1

Point 3

Fig 14 Model ten sion component forcendashdisplacement curve including force reversal

For

ce

Point 5 (ultimate strength)

Unloading curve

Intersection point

Reference point 1Displacement

(FA DA)

Fig 15 Unloading at changing temperatures

For

ce

DisplacementReference point 1

T1 gt T1F1 gt F2D2 gt D1

F1 D1

F2 D2

T1

T2

Fig 11 Assembly of the individual tension components to tension bolt row

F

D

Unloading curve

F

D

Unloading curve

F

D

+ + =

F

D

Comp 1 Comp 3Comp 2

Fig 12 Effective forcendashdisplacement curve of a typical tension bolt row

Displacement

Residual compressionContact of the beam weband column

Plastically deformed end-plate and column flange pushed back until centres are in contact

Tension

Bolt plastic deformation Tensionreduced to zero end-plate andcolumn flange not in contact

For

ce


componentrsquos current permanent defor-mation is that saved from the previous step and the permanent deformation is updated at the end of each step Figure 15 shows how this concept is implemented Reference point 1 is updated at the end of the step at tem-perature T1 mo ving to the next step (temperature T2) the unloading curve is plotted on the basis of the compo-nentrsquos new FndashD curve Therefore the new unloading curve will be located by starting from a point on the new load-ing curve and passing through refer-ence point 1 Finally the effective FndashD curve is formed for this temperature

Analytical Implications

Because of the nature of conven-tional quasi-static analysis an analy-

Fig 17 Staticndashdynamic progressive collapse modelling of a two-dimensional frame with five-row end-plate connections (a) ini-tial detachment of beam connections (b) column buckling at higher temperature (c) component forces up to connection failure (d) connection rotations and column displacements

1000200 400 600 8000

Displacement (mm)

Displacement of top of column C1

0 005 01 015 02 025 03 035 04

Rotation (rad)

Beam end rotation at J1

700

600

500

400

300

200

100

Tem

pera

ture

(degC

)

Forces in component (kN)

Top bolt rowSecond bolt rowThird bolt rowFourth bolt rowBottom bolt row

800

700

600

500

400

300

200

100

0

Tem

pera

ture

(degC

)

800

1209060300

Component fracture

J1

C1

J1

C1

(c) (d)

(a)

(b)

Fig 16 Principles of the staticndashdynamic analysis

Load(or temperature)

Deflection

Stab

le r

egio

n

Stable regionUnstable region

DynamicCritical

sis of a structure in a fire which includes component-based connection elements can only trace the behav-iour of a connection up to the point where its first component fails In real-ity a connection may either be able to regain its capacity after the initial frac-ture of a component or the first fail-ure may trigger a cascade of failures of other components leading to complete detachment of the connected member This possibility should be considered in performance-based design when a structure is being tested for robust-ness If connections are to avoid the possibility of becoming detached from members this numerical model-ling must be capable of predicting the sequence of failures of components rather than simply the first loss of sta-bility A numerical procedure in which

the whole behaviour from first insta-bility to total collapse can be modelled effectively has recently56 been devel-oped in Vulcan

The Vulcan model combines alternate static and dynamic analyses in order to use both to best advantage Static anal-ysis is used to follow the behaviour of the structure at changing temperatures until instability happens beyond this point an explicit dynamic procedure is activated to track the motion of the system until stability is regained The process is illustrated schematically in Fig 16 When combined with the par-allel development of general compo-nent-based connection elements which has been described this procedure can effectively track the behaviour of con-nections from the initial fracture of a component via the failure of suc-cessive bolt rows to final detachment from the column Even then if the remaining structure can carry the load-ing with its current temperature dis-tribution the analysis can re- stabilize once again In fact the analysis of a simple frame model depicted in Fig 17 carries on beyond connection frac-ture row-by-row includ ing complete detachment of the heated beam until the final structural collapse of the frame occurs due to column buckling at a higher temperature

Experiments on Connections under Combined Forces

Between 2005 and 2008 the Universities of Sheffield and Manchester collabo-


Fig 18 Schematic of electric furnace and test set-up for multi-directional loading tests

Load jack

Reaction frame

Electrical furnace

Macalloy bars

Testedconnection

Reaction frame

Support beam

αFurnacebar Link

bar Jackbar

CameraCamera

Fig 19 Force-rotation plots for 10 mm end-plate connections

0

40

80

120

160

200

240

280

1 3 5 7 9 Rotation (deg)

For

ce (

kN)

45deg Load angle35deg Load angle

55deg Load angle

Fig 20 Effect of end-plate thickness

0

20

40

60

80

100

120

140

0 3 6 9 12 15Rotation (deg)

For

ce (

kN)

tp = 10 mm tp = 8 mm

tp = 15 mm

550degC

Fig 21 Effect of number of bolt rows

0

50

100

150

200

250

300

350

0 3 6 9 12Rotation (deg)

For

ce (

kN)

2 Bolt rows

3 Bolt rows20degC

550degC

rated in a research programme investi-gating the capacity and ductility of steel connections at elevated temperatures The investigation adopted a test set-up in which the connections were sub-jected to a combination of tension and shear forces as well as high rotations Moments and rotations were gener-ated at the connections due to the lever arm of the applied force In total four types of connection were studied flush end-plates flexible end-plates fin-plates and web cleats The objective of these tests was to provide carefully monitored data on the behaviour and progressive failure of rea listic connec-tions under conditions similar to those in framed structures in a fire so that component models and component-based elements could be tested and developed In all cases a UC254 times 89 section was used for the column and the beam specimens were all UB305 times 165 times 40

Semi-Rigid Conn ections Flush End-Plates

The momentndashrotation characteristics of flush end-plate connections have been investigated2457 previously at ambient and elevated temperatures Normal calculation of their tying capacity assumes that the connection

is subjected to pure tension and that each bolt row contributes fully to its resistance This is obviously impossible in practice Coexisting actions may overload individual fasteners so that all the bolt rows do not reach their maximum resistance at the same time if their behaviour is not ductile enough and this may cause an ldquounzippingrdquo fail-ure Most tests used three bolt rows but for two tests the middle bolt row was removed The c onnections were tested at three different combinations of shear and tying force corresponding to different angles α in Fig 18

The forcendashrotation relationships for the t ests using 10 mm end-plates are shown in Fig 19 At 550degC the test at 45deg fai led because of thread strip-ping from the nuts subsequently two nuts were used on each bolt to prevent thread stripping The resistance of the connection reduced rapidly with the increase in temperature The load angle had some effect on the overall connec-tion resistance but not on the failure mode Figure 20 shows the main effect of end-plate thickness on the response of the connection a thick end-plate enhances resistance but significantly reduces ductil ity Figure 21 compares two tests with three rows against two tests with two rows removing the mid-dle bolt row clearly reduces the resis-tance but is also seen from the results at 550degC to reduce the ductility

For the tests with a 10 mm thick end-plate and three bolt rows two failure modes were observed At 20 and 450degC failure was controlled by end-plate

fracture Fig 22(a) shows an example after a test at 450degC At 550 and 650degC failure was controlled by the very duc-tile bolt extension characteristics as shown in Fig 22(b) For the 15 mm thick end-plate the failure was unsur-prisingly controlled by the bol ts

Simple Connections

Similar tests have been performed on flexible end-p late fin-plate and web cleat connections which are commonly used simple connections designed according to the ldquoGreen Bookrdquo58 rec-ommendations The responses of these simple connections are compared with the flush end-plate connection in Fig 23

All the flexible end-plate connections that were te sted59 failed because of th e fracture of the end-plate in t he heat-affected zone adjacent to the welds to the beam web with relatively low rotational capacity at high tem pera-tures All the tested fin-plate connec-tions60 failed because of shear fracture of their bolts Bolt clearance at holes allowed the connection a rotation of up to 4deg before the bearing surfaces were in contact This gave them a rota-tion capacity slightly better than that of flexible end-plates The ldquoGreen Bookrdquo notes that bolt shear fracture can be avoided by limiting the thick-ness of the bearing plate to less than half of the bolt diameter This proved to be inadequate at high temperatures Other tests using grade 109 bolts suc-cessfully changed the failure mode to block shear fracture of the beam web and increased the rotation capacity by about 3deg at ambient temperature However this benefit is not seen at high temperatures since the failure is again due to shear fracture of the bolts

The web cleat connections61 failed in a more complex fashion At ambient temperature the bolt head punched through the angle connected to the column flange At 450 and 550degC the angle fractured close to its heel at a significantly smaller deformation than at ambient temperature At 650degC the failure of the connection was by shear fracture of the b olts through the beam web At all temperatures web cleat connections showed high rotation capacity due to the ldquostraight-eningrdquo of the angle cleats With the increase in rotation the load capacity increased steadily giving the web cleat connections a significantly higher ulti-mate resistance than the other simple connections


Fig 24 Comparison of test results at 20 450 550 and 650degC with component-based modelling for α = 35deg

0

50

100

150

200

250

0 3 6 9 12 15 18 21Rotation (deg)

Tot

al fo

rce

(kN

)

Test

Component model20degC

450degC

550degC

650degC

Fig 22 (a) Failure of end-plate connection at 450degC (b) failure of end-plate connection at 550degC 650degC

(a)

(b)

Fig 23 Comparison of the behaviour of different connection types

0

50

100

150

200

250

300

0 4 8 12 16 20

Rotation (deg)

0 4 8 12 16 20

Rotation (deg)

Rotation (deg) Rotation (deg)

20degC0

0 3 6 9 12 15

For

ce (

kN)

20

40

60

80

100

120

140

160

180

450degC

For

ce (

kN)

00

100

10

20

30

40

50

60

70

80

90

5 10 15

550degC

For

ce (

kN)

Flush epFlexible epWeb cleatFin-plate




0

5

10

15

20

25

30

35

40

45

650degC

gated the behaviour and robustness of practical connections between steel or composite beams and two types of composite columns in a firemdashconcrete-filled tubes and partially encased (flange-infilled) H-sections The experimental investigation on flush end-plate and reverse channel connections at elevated temperatures and the deve lopment of componen t-based models for such connections have been carried out The test set-up shown in Fig 18 was reused to conduct constant temperature tests with load-i ng under displacement control until fracture occurred Figure 25 shows two typical specimen configurations

It was found that the reverse channel connections provided at least three times more rotation capacity t han the equivalent flush end-plate connec-tions tested at the same temperature alth ough with comparable ultimate strength Th e main failure modes of the reverse channel connections were frac-ture of the reverse channel web bolt heads punching through bolt holes and tensile fract ure of bolts In no tes ts was there noticeable deformation of the concrete-filled tubular (CFT) columns or of the steel beams Neither was any damage found to the connection welds All reverse channels experienced large plastic deformation (Fig 26) before failure occurred showing clearly the very high ductility achieved

On the basis of the experiments and the FE studies the active components for reverse channel connections have been identified these are illustrated in Fig 27 Component char acteristics developed previously556364 have been used where these components (eg bolts in tension) exist component models for reverse channels themselves were not available and so have been developed in COMPFIRE These component models have been integrated into the component-based connection element

Components have been characterized for all the connection types tested An example of the simulation of the Sheffield tests for the tests at ambi-ent temperature and at three elevated temperatures is shown in Fig 24 The component-based model gives a

reasonable repre sentation of the test behaviour

Other Connection Types

The Sheffield team participated in the European collaborative project COMPFIRE62 This project investi-


Fig 25 Test specimens (a) flush end-plate connection (b) reverse channel connection

UC 254 times254 times 89

UB305 times 165 times 40

10 mm fillet weld alongchannel length to tubeView 1-1 View 1-1

UB 305 times 165 times 40

UKPFC 200 times 90 times 30CHS 2445 times 8

1 111

(a) (b)

of beam

300

300

100

320

100

20 20 7055 20 20325 70

400

320

CL of beamCL

Fig 26 Typical failure of tube-cut reverse channel connections

Fig 27 Active components of reverse channel joints

Reverse channel in compression

Reverse channel in bending

Endplate in bending

Bolt in tension

M

Fig 28 Validation of the integrated com-ponent-based connection element against test data

160

140 CIDECT

AISC

Bolt pullout model

Connection element

Test

120

100

80

60

40

20

00 5 10

Rotation (deg)

Forc

e (k

N)

15 20 25

in Vulcan Figure 28 shows one exam-ple used to test the model against the COMPFIRE isolated joint tests

Conclusion

The response of structural frames subject to fire is highly dependent on

the behaviour of their joints During initial heating compressive forces are generated in the beam-to-column con-nections due to the restrained thermal expansion of the beams Some connec-tions can fail due to this force which has been suggested as the cause of failure of 7 World Trade48 As temperatures

rise further the compression is pro-gressively reduced by sagging deflec-tion of the beam and by degradation of material strength and stiffness At very high temperatures the beam may have lost nearly all its bending stiffness and experiences very large deflection At this stage the beam actually hangs in catenary tension between its end connections and whether the connec-tions have sufficient ldquotyingrdquo capacity determines whether they will fracture The ductile design of connections is important because the connection forces both in compression and in catenary tension can be reduced con-siderably if the connections themselves can deform and accommodate the end movements of the beams

It is essential to understand the behav-iour of connections in order to predict the global frame response to fire When modelling connections in an extensive building frame it is nearly impos-sible to model them in detail due to the complexity of their geometry and behaviour Instead they are usually oversimplified as either pinned or rigid which leads to unrepresentative results It has been found that a com-ponent-based approach can provide a sufficiently accurate and practical solution to the problem of modelling connections in a fire Previously com-ponent-based models were developed mainly to model rotational charac-teristics for the ambient temperature design of end-plate connections for semi-rigid frames but they are ideal for including normal force and deformation as part of a linked non-linear structural model Through a series of research projects the behaviour of most components of a range of con-nection types tested has been repr e-sented in simplified high-temperature non-linear spring models


[19] Leston-Jones LC The Influence of Semi-rigid Connections on the Performance of Steel Framed Structures in Fire PhD Thesis University of Sheffield 1997

[20] Spyrou S Davison JB Burgess IW Plank RJ Experimental and anal ytical investigation of the tension zone component within a steel joint at elevated temperatures J Const Steel Res 2004 60(6) 867ndash896

[21] Spyrou S Davison JB Burgess IW Plank RJ Experimental and anal ytical investigation of the compression zone component within a steel joint at elevated temperatures J Const Steel Res 2004 60(6) 841ndash865

[22] Simotildees da Silva L Santiago A Vila Real P A component model for the behaviour of steel joints at elevated tem peratures J Const Steel Res 2001 57 1169ndash1195

[23] Al-Jabri KS Component-based model of the behaviour of flexible end-plate connections at elevated temperatures Compos Struct 2004 66 215ndash221

[24] Al-Jabri KS Burgess IW Plank RJ Spring-stiffness model for flexible end-plate bare-steel joints in fire J Const Steel Res 2005 61 1672ndash1691

[25] Luciano de Lima RO Simotildees da Silva L Vellasco PCG Andrade SAL Experime ntal analysis of extended end-plate beam-to-column joints under bending and axial force in Proc Eurosteel Coimbra Portugal 2002

[26] Luciano de Lima RO Simotildees da Silva L Vellasco PCG Andrade SAL Experimental evaluation of extended endplate beam-to- column joints subjected to bending and axial force Engnr Struct 2004 26 1333ndash1347

[27] Jaspart J-P Recent advances in the field of steel joints Column bases and further con-figurations for beam-to-column joints and beam splices University of Liegravege Department MSM Belgium 1997

[28] Jaspart J-P General report session on con-nections J Const Steel Res 2000 55 69ndash89

[29] Cerfontaine F Jaspart J-P Analytical Study of the Interaction Between Bending and Axial Force on Bolted Joints in Proc Eurosteel Coimbra Portugal 2002

[30] Wald F Svarc M Experiments with End Plate Joints Subject to Moment and Normal Force Contributions to Experi mental Inves-tigation of Engineering Materials and Structures CTU Reports No 2-3 Prague 2001

[31] Luciano de Lima RO Simotildees da Silva L Vellasco PCG Andrade SAL Experimental analysis of extended end-plate beam-to-col-umn joints under bending and axial force in Proceedings of the third European conference on Steel Structiures Coimbra Portugal 2002

[32] Luciano de Lima RO Simotildees da Silva L Vellasco PCG Andrade SAL Experimental evaluation of extended endplate beam-to-col-umn joints subjected to bending and axial force Engnr Struct 2004 26 1333ndash1347

[33] Liu TCH Fahad MK Davies J Experimental investigation of behaviour of axially restrained steel beams in fire J Const Steel Res 2002 58 1211ndash1230

[34] Mesquita LMR Piloto PAG Vaz MAP Vila Real PMM Experimental and numerical

References

[1] Bailey CG Structural fi re design core or spe-cialist subject Struct Eng 2004 82(9) 32ndash38

[2] CEN EN 1993-1-2 2005 Eurocode 3 Desi gn of Steel Structures Part 12 Structural Fire Design European Committee for Standardisation Brussels 2005

[3] Franssen J-M Numerical Determination of 3D Temperature Fields in Steel Joints In 2nd International Workshop on Structures in Fire Christchurch New Zealand 2002

[4] Nethercot DA Frame structures global performance static and stability behaviou rmdash general report J Constr Steel Res 2000 55(1ndash3) 109ndash124

[5] British Standards Institution BS 476 Method for Determination of Fire Resistance of Elements of Construction Part 20 BSI London 1990

[6] SCI Investigation of Broadgate Phase 8 Fire Structural Fire Engineering Steel Construction Institute Ascot UK 1991

[7] Al-Jabri KS Modelling of beam-to- column connections at elevated temperature in High performance structures and materials II Brebbia C W W Ed 2004 pp 319ndash328

[8] Kruppa J Reacutesistance en feu des assemblages par boulons Centre Technique Industriel de la Construction Meacutetallique St Reacutemy les Chevreuse France 1976

[9] British Steel Corporation The Performance of BeamColumnBeam Connections in the BS 5950 Part 8 Fire Test British Steel (Swinden Laboratories) Rotherham UK 1982

[10] Lawson RM Behaviour of steel beam-to-column connections in fire The Struct Engnr 1990 68(14) 263ndash271

[11] Leston-Jones LC Lennon T Plank RJ Burgess IW Elevated temper ature moment-rotation tests on steelwork connections Proc Instn Civ Engrs Structs Bldgs 1997 122 410ndash419

[12] Davison JB Kirby PA Nethercot DA Rotational stiffness characteristics of steel beam to column conne ctions J Const Steel Res 1987 18 17ndash54

[13] Al-Jabri KS Lennon T Burgess IW Plank RJ Behaviour of steel and composite beam-column connections in fire J Const Steel Res 1998 46(1ndash3) 308ndash309

[14] Al-Jabri KS Burgess IW Lennon T Plank RJ Moment-rotation-temp erature curves for semi-rigid joints J Const Steel Res 2005 61 281ndash303

[15] Zoetemeijer P A design method for the tension side of statically loaded bolted beam-to-column connections Heron 1974 20 1ndash59

[16] Tschemmernegg F Tautschnig A Klein H Braun C Humer C Zur Nachgiebigkeit von Rahmenknoten ndash Teil 1 (Semi-rigid joints of frame structures Vol 1) Stahlbau 1987 56 299ndash306

[17] COST Project C1 Semi-Rigid Behaviour Steel and Composite Group C1WD298-03 Innsbruck Austria 1998

[18] CEN EN 1993-1-8 200 5 Eurocode 3 Design of Steel Structures Part 1-8 General Rules Design of Joints European Committee for Standardisa-tion Brussels 2005

Components so far characterized have been shown to predict the connection behaviour with satisfactory accuracy The component-based model has been assembled as a connection element in the Vulcan software and this develop-ment has been made in parallel with implementation of a staticdynamic solution pro cess This combination allows the behaviour of a building frame to be modelled throughout the course of a fire so that progressive failures of parts of connections do not cause a premature termination of the analysis due to numerical instabil-ity This kind of analysis is necessary for true performance-based design of framed buildings against fire so that potential disproportionate collapse can be predicted and prevented by adjusting the design of the structure including that of its connections

The research so far has neglected detailed testing and validation in the initial heating phase which causes axial compression in beams and their connections However some types of connection (the more obvious being fin-plates and web cleats) can either fracture components completely or damage them severely in thi s phase and research work remains to be done on this phase of behaviour Before the component-based approach or generalized design rules can be rec-ommended for adoption the perfor-mance throughout the whole cycle of compressive-tensile displacement combined with rotation needs to be investigated both in the context of whole connections and their compo-nents at different temperatures The continuity of slabs and their rebar over the top of internal beamndash column connections clearly increases the rotational stiffness of a connection However in a region of high local-ized rotation it may fracture relatively early in the initial heating phase when the rotation is caused mainly by ther-mal bowing This is being investigated in a current project

Acknowledgements

The research leading to these results has received funding from various sources These include four major tranches of support from the Engineering and Physical Sciences Research Council of the United Kingdom and one from the European Communityrsquos Research Fund for Coal and Steel (Grant Agreement RFSR-CT-2009-00021) The authors wish to gratefully acknowledge the contribution to their work made by these bodies


Connections at Elevated Temperatures PhD the-sis University of Sheffield 2006

[56] Sun RR Huang Z Burgess IW Progressive collapse analysis of steel structures under fire conditions Engnr Struct 2012 34 400ndash413

[57] Aggarwal A K Comparative tests on end-plate beam-to-column connections J Construct Steel Res 1994 30 151ndash175

[58] SCIBCSA Joints in Steel Construction Simple Connections The Steel Construction Institute and British Constructional Steelwork Association London UK 2002

[59] Hu Y Davison JB Burgess IW Plank RJ Component modelling of flexible end-plate con-nections in fire Int J Steel Struct 2009 9 29ndash38

[60] Yu HX Burgess IW Davison JB Plank RJ Experimental investigation of the behaviour of fin plate connections in fire J Construct Steel Res 2009 65 723ndash736

[61] Yu HX Burgess IW Davison JB Plank RJ Tying capacity of web cleat connections in fire Part 1 test and finite element simulation Eng Struct 2009 31(3) 651ndash663

[62] RFCS COMPFIRE ndash Design of joints to composite c olumns for improved fire robust-ness Research Fund for Coal and Steel Grant agreement no RFSR-CT-2009-00021 European Commission Brussels 2009

[63] Spyrou S Development of a Component-Based Model of Steel Beam-to-Column Joints at Elevated Temperatures PhD Thesis Sheffield University of Sheffield 2002

[64] Yu H Bur gess IW Davison JB Plank RJ Development of a yield-line model for endplate connections in fire J Construct Steel Res 2009 65(6) 1279ndash1289

[45] Gillie M Usmani AS Rotter JM A struc-tural analysis of the first Cardington test J Const Steel Res 2001 57 581ndash601

[46] FEMAASCE World Tra de Centre Building Performance Study 2002

[47] IStructE Safety in Tall Buildings and Other Buildings with Large Occupancy IStructE London 2002

[48] NIST Prevention of Progressive Collapse Report on July 2002 National Workshop and Recommendations for Future Efforts Multi Hazard Mitigation Council of NIST Washington DC 2003

[49] Arup Ltd A Scoping Studymdashthe Building Regulations Post September 11 Arup 2003

[50] Ding J Behaviour of Restrained Concrete Filled Tubular Columns and Their Joints in Fire University of Manchester UK 2007

[51] Huang Z Burgess IW Plank RJ Modelling of six full-scale fire tests on a composite building Struct Engnr 2002 80(19) 30-37

[52] Franssen J-M SAFIR A thermal structural program modelling structures under fire Engnr J Am Inst Steel Constuct 2005 42(3) 123ndash158

[53] Block FM Burgess IW Davison JB Plank RJ The development of a component-based connection element for endplate connections in fire In 4th International Workshop on Structure in Fire Aveiro Portugal 2006

[54] Masing G Zur Heynschen Theorie der Verfestigung der Metalle durch verborgen elas-tische Spannungen Wiss Veroffentl aus dem Siemens-Konzern 1923 31 231ndash239

[55] Block FM Development of a Component-Based Finite Element for Steel Beam-to-Column

research on the critical temperature of laterally unre strained steel I beam J Const Steel Res 2005 61 1435ndash1446

[35] Simotildees da Silva L Santiago A Vila Real P Moore DB Behaviour of steel joints under fire loading Steel Compos Struct 2005 5(6) 485ndash513

[36] Wang Y Dai X Bailey C An experimental study of relative structural fire behaviour and robustness of different types of steel joint in restrained steel frames J Const Steel Res 2011 67(7) 1149ndash1163

[37] Foster SJ Chladna M Hsieh Y-C Burgess IW Plank RJ Thermal and structural behaviour of a full-scale composite building subject to a severe compartment fire Fire Safety J 2007 42 183ndash199

[38] Armer GST Moore DB Full-scale test-ing on complete multi-storey structures Struct Engnr 1994 72(2) 30ndash31

[39] Moore DB Lennon T Fire engineering design of steel structures Prog Struct Engnr Mater 1997 1(1) 4ndash9

[40] Lennon T Cardington Fire Tests Survey of Damage to Eight Storey Building Building Research Establishment Garston UK 1997

[41] Simms WI The Cardington Fire Tests SCIIStructE London 1998

[42] Al-Jabri KS Hago AW Towards a rational approach to the design of steel-framed build-ings in fire in 9th Arab Structural Engineering Conference United Arab Emirates 2003

[43] Wald F Simotildees da Silva L Moore DB Santiago A Experimental behaviour of steel joints under natural fire in ECCS-AISC Workshop 2004

[44] Al-Jabri KS Burgess IW Lennon T Plank RJ The performance of frame connections in fire Acta Polytechnica 1999 39(5) 65ndash75

Sign up for an Institutional Subscription 2013

Structural Engineering International (SEI) IABSE Congress and Symposium Reports

SEI current + back issues (since 1991) CHF 630

IABSE Congress and Symposium Reports (since 2000) and SEI current + back issues (since 1991) CHF 830

‐ Online access for the entire Institution via site license‐ Subscription facility is available with all major Subscription Agencies

SEI includes over 1500 papers Reports over 3500 papers

Subscribe nowwwwiabseorgjournalseisubscription

Option 2

Option 1


joints that are assumed to be pinned at ambient temperature can provide sig-nificant levels of strength and stiffness at elevated temperatures especially as their temperatures remain lower than those of the connected beams They can therefore reduce significantly the deflections of these beams at any given stage of a fire which increases their fire resistance in conventional5 fur-nace testing It is clear that the internal moments in beams can be redistrib-uted to adjacent cooler members as was illustrated in the Broadgate fire report6

Experimental investigations con-ducted on the performance of steel joints at elevated temperatures are relatively recent and rather limited in number mainly because of the high cost of fire tests and the limitations on the size of the furnace used The joint tests conducted have primarily focused on establishing momentndashrotation rela-tionships of isolated joints However even on these terms an experimental fire test is incapable of extrapolation to details other than those used in the actual test or to other fire timendashtem-perature relationships Results from fire tests of i solated joints provide important fundamental data on the behaviour of joints but do not truly reflect the actual behaviour of joints in buildings in the event of a fire due to the absence of structural continu-ity Thus in recent years the numeri-cal simulation of joints and structural frames subjected to fire conditions has been the basis of extensive research7 This paper attempts to trace the devel-opment of research work that initially concentrated on investigating the rota-tional behaviour of beam-to-column joints at elevated temperatures This experimental work on isolated joints has le d to the development of various analytical approaches to predict the r otational behaviour of both bare-steel and composite joints and the effect of structural continuity on the perfor-mance of beams supported by realistic joints

Rotational Behaviour of I solated Joints in a Fire

The first experimental fire tests on joints were conducted by Centre Technique Industriel des Constructions Metalliques (CTICM)8 on six types ranging from ldquoflexiblerdquo to ldquorigidrdquo The primary purpose of these tests was to investigate the performance of high-strength bolts at elevated temperatures

Introduction

The mechanical properties of steel structures de grade rapidly in a fire because of the reduction in both the stiffness and strength of the material at high temperatures Using applied fire protection (covering the exposed steel with a prescribed thickness of an insulating material) remains the most common way of satisfying the fire resistance requirement for a steel-framed structure However it is more rational to consider fire as an addi-tional load case1 just as designers routinely consider wind or earthquake hazards and to design the structure with appropriate protection to be acceptable at the fire limit state rather than designing the structure for other limit states and then applying retro-spective fire protection Following the recognition of this logic interest has grown in understanding the behaviour of structural elements in a fire both in isolation and as part of the whole building Experimental and theoretical research has made considerable prog-ress towards the goal of considering the fire limit in the normal analytical design process

All structural members exposed to fire heat up but with different rates of tem-perature rise Joints in a steel-framed building tend to heat more slowly than the material within the free span of the beams that they connect because of the additional mass of material (bolts plates angles etc) in a shielded loca-tion (ie usually beneath a composite floor) and a relatively small surface area EN 199 3-1-2 20052 suggests sim-plified joint temperatures of between 62 and 88 of that in the beam lower flange temperature at mid-span Alternatively the temperature distri-bution may be calculated on the basis of the exposed area volume ratio of each joint component (plate angle bolt etc) Franssen3 has suggested that this method may be non-conservative

Traditionally the design of steel-framed structures has assumed that the connection between a beam and a column is either rigid (implying com-plete rotational continuity) or pinned (implying that there is no moment transfer) However actual connection behaviour exhibits characteristics over a wide spectrum between these two limits connections regarded as pinned generally possess some rotational stiffness whilst ldquorigidrdquo connections display some rotational flexibility4 It was realized in the late 1980s that

and no indication of the performance of the joints was reported Two tests were carried out by British Steel9 in 1982 on a ldquorigidrdquo moment-resisting joint which sustained significant deformation during a fire The first tests to assess the structural continu-ity afforded by beam-to-column joints at elevated temperatures were car-ried out by Lawson10 with the aim of developing a design approach for steel beams taking into account the rota-tional restraint provided by end-plate or web angle joints The tests dem-onstrated the strength of these joints in a fire and showed that significant moments (up to two thirds of their ambient temperature design moment capacity) could be sustained in fire conditions Lawson proposed simple rules for designing simply supported beams in a fire taking into account the moment transferred to columns via joints in fire conditions Although the test results provided insufficient data to describe the momentndashrotation char-acteristics of the joints they provided essential information for the early attempts at joint modelling

The first furnace tests intended to develop momentndashrotation relation-ships for flush end-plate joints at elevated temp eratures were con-ducted by Leston-Jones et al 11 Eleven tests were carried out on small-scale specimens including two at ambient temperature for both bare-steel and composite joints The results demon-strated that both the initial stiffness and moment capacity of the joints decreased with increasing tempera-ture with a significant reduction in capacity for temperatures in the range of 500 to 600degC These tests provided useful data for connection model-ling but for a limited range of details employing relatively small section sizes for comparison with earlier joint testing work at ambient temperature by Davison et al12 A further series of elevated temperature joint tests was conducted by Al-Jabri et al13 to study the influence of details such as mem-ber size end-plate type and thic kness and composite slab characteristics on the joint response in a fire For each joint a series of tests was conducted at a constant moment but with increas-ing furnace temperature14

Prediction of Joint Characteristics Using Component-Based Models

The component method was devel-oped following the suggestion1 516 in the 1970s that it represented in a prac-





Kt

Kc

ZM

P


Ft

Fc

Ft4

Ft3

Ft2

Ft1

c1

t1 t2 t3 t4

c2c3c4c t

Fc1

Fc2

Fc3

Fc4

Kt1

Kc1

Kc2

Kc3

Kc4

Kt2

Kt3

Kt4


900

800

700

600

500

400

300

200

100


Bea

m fl

ange

tem

pera

ture

(degC

)

60 70 80 90 100 110 120 130

5

1015 20

25

Connection

Model

Test

Moment (kN m)



















(a) (b)














400

(a) (b)

200

0

minus200

minus400

minus600

minus800

0 200 400 600 800 1000 1200

Temperature (degC)

Axi

al fo

rce

(kN

)

400

200

0

minus200

minus400

minus600

minus800

Axi

al fo

rce

(kN

)



Tension

Compression

Heating

Cooling

0 200 400 600 800 1000 1200

Temperature (degC)



Tension

Compression

Heating

Cooling



Tension C omponent









42 3 1 3

5 Mi Ni ui

Nj uj

Vj wj

jiVi wi

(a) (b)

k1

k2

k3

i Mj j

0 mmw

u

l c2

l c1

lT2

lT3

lT1


0

100

200

300

400

500

600

700

800

900

0 10 20 30 40 50 60 70 80 90 100


Stee

l tem

pera

ture

(degC

)









11 2










D = f(F) (1)


(DA ndash D)

_________ 2 = f (

(FA ndash F) ________

2 ) (2)





Displacement

For

ce

Point 2

Point 1

Point 3


For

ce


Unloading curve

Intersection point


(FA DA)


For

ce



F1 D1

F2 D2

T1

T2


F

D

Unloading curve

F

D

Unloading curve

F

D

+ + =

F

D

Comp 1 Comp 3Comp 2


Displacement



Tension


For

ce






1000200 400 600 8000

Displacement (mm)


0 005 01 015 02 025 03 035 04

Rotation (rad)


700

600

500

400

300

200

100

Tem

pera

ture

(degC

)



800

700

600

500

400

300

200

100

0

Tem

pera

ture

(degC

)

800

1209060300

Component fracture

J1

C1

J1

C1

(c) (d)

(a)

(b)



Deflection

Stab

le r

egio

n


DynamicCritical








Load jack

Reaction frame

Electrical furnace

Macalloy bars

Testedconnection

Reaction frame

Support beam

αFurnacebar Link

bar Jackbar

CameraCamera


0

40

80

120

160

200

240

280


For

ce (

kN)


55deg Load angle


0

20

40

60

80

100

120

140


For

ce (

kN)


tp = 15 mm

550degC


0

50

100

150

200

250

300

350


For

ce (

kN)

2 Bolt rows

3 Bolt rows20degC

550degC








Simple Connections






0

50

100

150

200

250

0 3 6 9 12 15 18 21Rotation (deg)

Tot

al fo

rce

(kN

)

Test


450degC

550degC

650degC


(a)

(b)


0

50

100

150

200

250

300

0 4 8 12 16 20

Rotation (deg)

0 4 8 12 16 20

Rotation (deg)


20degC0

0 3 6 9 12 15

For

ce (

kN)

20

40

60

80

100

120

140

160

180

450degC

For

ce (

kN)

00

100

10

20

30

40

50

60

70

80

90

5 10 15

550degC

For

ce (

kN)





0

5

10

15

20

25

30

35

40

45

650degC















1 111

(a) (b)

of beam

300

300

100

320

100

20 20 7055 20 20325 70

400

320

CL of beamCL





Endplate in bending

Bolt in tension

M


160

140 CIDECT

AISC

Bolt pullout model

Connection element

Test

120

100

80

60

40

20

00 5 10

Rotation (deg)

Forc

e (k

N)

15 20 25


Conclusion






















References





















Acknowledgements










































Option 2

Option 1





Kt

Kc

ZM

P


Ft

Fc

Ft4

Ft3

Ft2

Ft1

c1

t1 t2 t3 t4

c2c3c4c t

Fc1

Fc2

Fc3

Fc4

Kt1

Kc1

Kc2

Kc3

Kc4

Kt2

Kt3

Kt4


900

800

700

600

500

400

300

200

100


Bea

m fl

ange

tem

pera

ture

(degC

)

60 70 80 90 100 110 120 130

5

1015 20

25

Connection

Model

Test

Moment (kN m)



















(a) (b)














400

(a) (b)

200

0

minus200

minus400

minus600

minus800

0 200 400 600 800 1000 1200

Temperature (degC)

Axi

al fo

rce

(kN

)

400

200

0

minus200

minus400

minus600

minus800

Axi

al fo

rce

(kN

)



Tension

Compression

Heating

Cooling

0 200 400 600 800 1000 1200

Temperature (degC)



Tension

Compression

Heating

Cooling



Tension C omponent









42 3 1 3

5 Mi Ni ui

Nj uj

Vj wj

jiVi wi

(a) (b)

k1

k2

k3

i Mj j

0 mmw

u

l c2

l c1

lT2

lT3

lT1


0

100

200

300

400

500

600

700

800

900

0 10 20 30 40 50 60 70 80 90 100


Stee

l tem

pera

ture

(degC

)









11 2










D = f(F) (1)


(DA ndash D)

_________ 2 = f (

(FA ndash F) ________

2 ) (2)





Displacement

For

ce

Point 2

Point 1

Point 3


For

ce


Unloading curve

Intersection point


(FA DA)


For

ce



F1 D1

F2 D2

T1

T2


F

D

Unloading curve

F

D

Unloading curve

F

D

+ + =

F

D

Comp 1 Comp 3Comp 2


Displacement



Tension


For

ce






1000200 400 600 8000

Displacement (mm)


0 005 01 015 02 025 03 035 04

Rotation (rad)


700

600

500

400

300

200

100

Tem

pera

ture

(degC

)



800

700

600

500

400

300

200

100

0

Tem

pera

ture

(degC

)

800

1209060300

Component fracture

J1

C1

J1

C1

(c) (d)

(a)

(b)



Deflection

Stab

le r

egio

n


DynamicCritical








Load jack

Reaction frame

Electrical furnace

Macalloy bars

Testedconnection

Reaction frame

Support beam

αFurnacebar Link

bar Jackbar

CameraCamera


0

40

80

120

160

200

240

280


For

ce (

kN)


55deg Load angle


0

20

40

60

80

100

120

140


For

ce (

kN)


tp = 15 mm

550degC


0

50

100

150

200

250

300

350


For

ce (

kN)

2 Bolt rows

3 Bolt rows20degC

550degC








Simple Connections






0

50

100

150

200

250

0 3 6 9 12 15 18 21Rotation (deg)

Tot

al fo

rce

(kN

)

Test


450degC

550degC

650degC


(a)

(b)


0

50

100

150

200

250

300

0 4 8 12 16 20

Rotation (deg)

0 4 8 12 16 20

Rotation (deg)


20degC0

0 3 6 9 12 15

For

ce (

kN)

20

40

60

80

100

120

140

160

180

450degC

For

ce (

kN)

00

100

10

20

30

40

50

60

70

80

90

5 10 15

550degC

For

ce (

kN)





0

5

10

15

20

25

30

35

40

45

650degC















1 111

(a) (b)

of beam

300

300

100

320

100

20 20 7055 20 20325 70

400

320

CL of beamCL





Endplate in bending

Bolt in tension

M


160

140 CIDECT

AISC

Bolt pullout model

Connection element

Test

120

100

80

60

40

20

00 5 10

Rotation (deg)

Forc

e (k

N)

15 20 25


Conclusion






















References





















Acknowledgements










































Option 2

Option 1















(a) (b)














400

(a) (b)

200

0

minus200

minus400

minus600

minus800

0 200 400 600 800 1000 1200

Temperature (degC)

Axi

al fo

rce

(kN

)

400

200

0

minus200

minus400

minus600

minus800

Axi

al fo

rce

(kN

)



Tension

Compression

Heating

Cooling

0 200 400 600 800 1000 1200

Temperature (degC)



Tension

Compression

Heating

Cooling



Tension C omponent









42 3 1 3

5 Mi Ni ui

Nj uj

Vj wj

jiVi wi

(a) (b)

k1

k2

k3

i Mj j

0 mmw

u

l c2

l c1

lT2

lT3

lT1


0

100

200

300

400

500

600

700

800

900

0 10 20 30 40 50 60 70 80 90 100


Stee

l tem

pera

ture

(degC

)









11 2










D = f(F) (1)


(DA ndash D)

_________ 2 = f (

(FA ndash F) ________

2 ) (2)





Displacement

For

ce

Point 2

Point 1

Point 3


For

ce


Unloading curve

Intersection point


(FA DA)


For

ce



F1 D1

F2 D2

T1

T2


F

D

Unloading curve

F

D

Unloading curve

F

D

+ + =

F

D

Comp 1 Comp 3Comp 2


Displacement



Tension


For

ce






1000200 400 600 8000

Displacement (mm)


0 005 01 015 02 025 03 035 04

Rotation (rad)


700

600

500

400

300

200

100

Tem

pera

ture

(degC

)



800

700

600

500

400

300

200

100

0

Tem

pera

ture

(degC

)

800

1209060300

Component fracture

J1

C1

J1

C1

(c) (d)

(a)

(b)



Deflection

Stab

le r

egio

n


DynamicCritical








Load jack

Reaction frame

Electrical furnace

Macalloy bars

Testedconnection

Reaction frame

Support beam

αFurnacebar Link

bar Jackbar

CameraCamera


0

40

80

120

160

200

240

280


For

ce (

kN)


55deg Load angle


0

20

40

60

80

100

120

140


For

ce (

kN)


tp = 15 mm

550degC


0

50

100

150

200

250

300

350


For

ce (

kN)

2 Bolt rows

3 Bolt rows20degC

550degC








Simple Connections






0

50

100

150

200

250

0 3 6 9 12 15 18 21Rotation (deg)

Tot

al fo

rce

(kN

)

Test


450degC

550degC

650degC


(a)

(b)


0

50

100

150

200

250

300

0 4 8 12 16 20

Rotation (deg)

0 4 8 12 16 20

Rotation (deg)


20degC0

0 3 6 9 12 15

For

ce (

kN)

20

40

60

80

100

120

140

160

180

450degC

For

ce (

kN)

00

100

10

20

30

40

50

60

70

80

90

5 10 15

550degC

For

ce (

kN)





0

5

10

15

20

25

30

35

40

45

650degC















1 111

(a) (b)

of beam

300

300

100

320

100

20 20 7055 20 20325 70

400

320

CL of beamCL





Endplate in bending

Bolt in tension

M


160

140 CIDECT

AISC

Bolt pullout model

Connection element

Test

120

100

80

60

40

20

00 5 10

Rotation (deg)

Forc

e (k

N)

15 20 25


Conclusion






















References





















Acknowledgements










































Option 2

Option 1














400

(a) (b)

200

0

minus200

minus400

minus600

minus800

0 200 400 600 800 1000 1200

Temperature (degC)

Axi

al fo

rce

(kN

)

400

200

0

minus200

minus400

minus600

minus800

Axi

al fo

rce

(kN

)



Tension

Compression

Heating

Cooling

0 200 400 600 800 1000 1200

Temperature (degC)



Tension

Compression

Heating

Cooling



Tension C omponent









42 3 1 3

5 Mi Ni ui

Nj uj

Vj wj

jiVi wi

(a) (b)

k1

k2

k3

i Mj j

0 mmw

u

l c2

l c1

lT2

lT3

lT1


0

100

200

300

400

500

600

700

800

900

0 10 20 30 40 50 60 70 80 90 100


Stee

l tem

pera

ture

(degC

)









11 2










D = f(F) (1)


(DA ndash D)

_________ 2 = f (

(FA ndash F) ________

2 ) (2)





Displacement

For

ce

Point 2

Point 1

Point 3


For

ce


Unloading curve

Intersection point


(FA DA)


For

ce



F1 D1

F2 D2

T1

T2


F

D

Unloading curve

F

D

Unloading curve

F

D

+ + =

F

D

Comp 1 Comp 3Comp 2


Displacement



Tension


For

ce






1000200 400 600 8000

Displacement (mm)


0 005 01 015 02 025 03 035 04

Rotation (rad)


700

600

500

400

300

200

100

Tem

pera

ture

(degC

)



800

700

600

500

400

300

200

100

0

Tem

pera

ture

(degC

)

800

1209060300

Component fracture

J1

C1

J1

C1

(c) (d)

(a)

(b)



Deflection

Stab

le r

egio

n


DynamicCritical








Load jack

Reaction frame

Electrical furnace

Macalloy bars

Testedconnection

Reaction frame

Support beam

αFurnacebar Link

bar Jackbar

CameraCamera


0

40

80

120

160

200

240

280


For

ce (

kN)


55deg Load angle


0

20

40

60

80

100

120

140


For

ce (

kN)


tp = 15 mm

550degC


0

50

100

150

200

250

300

350


For

ce (

kN)

2 Bolt rows

3 Bolt rows20degC

550degC








Simple Connections






0

50

100

150

200

250

0 3 6 9 12 15 18 21Rotation (deg)

Tot

al fo

rce

(kN

)

Test


450degC

550degC

650degC


(a)

(b)


0

50

100

150

200

250

300

0 4 8 12 16 20

Rotation (deg)

0 4 8 12 16 20

Rotation (deg)


20degC0

0 3 6 9 12 15

For

ce (

kN)

20

40

60

80

100

120

140

160

180

450degC

For

ce (

kN)

00

100

10

20

30

40

50

60

70

80

90

5 10 15

550degC

For

ce (

kN)





0

5

10

15

20

25

30

35

40

45

650degC















1 111

(a) (b)

of beam

300

300

100

320

100

20 20 7055 20 20325 70

400

320

CL of beamCL





Endplate in bending

Bolt in tension

M


160

140 CIDECT

AISC

Bolt pullout model

Connection element

Test

120

100

80

60

40

20

00 5 10

Rotation (deg)

Forc

e (k

N)

15 20 25


Conclusion






















References





















Acknowledgements










































Option 2

Option 1



Tension C omponent









42 3 1 3

5 Mi Ni ui

Nj uj

Vj wj

jiVi wi

(a) (b)

k1

k2

k3

i Mj j

0 mmw

u

l c2

l c1

lT2

lT3

lT1


0

100

200

300

400

500

600

700

800

900

0 10 20 30 40 50 60 70 80 90 100


Stee

l tem

pera

ture

(degC

)









11 2










D = f(F) (1)


(DA ndash D)

_________ 2 = f (

(FA ndash F) ________

2 ) (2)





Displacement

For

ce

Point 2

Point 1

Point 3


For

ce


Unloading curve

Intersection point


(FA DA)


For

ce



F1 D1

F2 D2

T1

T2


F

D

Unloading curve

F

D

Unloading curve

F

D

+ + =

F

D

Comp 1 Comp 3Comp 2


Displacement



Tension


For

ce






1000200 400 600 8000

Displacement (mm)


0 005 01 015 02 025 03 035 04

Rotation (rad)


700

600

500

400

300

200

100

Tem

pera

ture

(degC

)



800

700

600

500

400

300

200

100

0

Tem

pera

ture

(degC

)

800

1209060300

Component fracture

J1

C1

J1

C1

(c) (d)

(a)

(b)



Deflection

Stab

le r

egio

n


DynamicCritical








Load jack

Reaction frame

Electrical furnace

Macalloy bars

Testedconnection

Reaction frame

Support beam

αFurnacebar Link

bar Jackbar

CameraCamera


0

40

80

120

160

200

240

280


For

ce (

kN)


55deg Load angle


0

20

40

60

80

100

120

140


For

ce (

kN)


tp = 15 mm

550degC


0

50

100

150

200

250

300

350


For

ce (

kN)

2 Bolt rows

3 Bolt rows20degC

550degC








Simple Connections






0

50

100

150

200

250

0 3 6 9 12 15 18 21Rotation (deg)

Tot

al fo

rce

(kN

)

Test


450degC

550degC

650degC


(a)

(b)


0

50

100

150

200

250

300

0 4 8 12 16 20

Rotation (deg)

0 4 8 12 16 20

Rotation (deg)


20degC0

0 3 6 9 12 15

For

ce (

kN)

20

40

60

80

100

120

140

160

180

450degC

For

ce (

kN)

00

100

10

20

30

40

50

60

70

80

90

5 10 15

550degC

For

ce (

kN)





0

5

10

15

20

25

30

35

40

45

650degC















1 111

(a) (b)

of beam

300

300

100

320

100

20 20 7055 20 20325 70

400

320

CL of beamCL





Endplate in bending

Bolt in tension

M


160

140 CIDECT

AISC

Bolt pullout model

Connection element

Test

120

100

80

60

40

20

00 5 10

Rotation (deg)

Forc

e (k

N)

15 20 25


Conclusion






















References





















Acknowledgements










































Option 2

Option 1










D = f(F) (1)


(DA ndash D)

_________ 2 = f (

(FA ndash F) ________

2 ) (2)





Displacement

For

ce

Point 2

Point 1

Point 3


For

ce


Unloading curve

Intersection point


(FA DA)


For

ce



F1 D1

F2 D2

T1

T2


F

D

Unloading curve

F

D

Unloading curve

F

D

+ + =

F

D

Comp 1 Comp 3Comp 2


Displacement



Tension


For

ce






1000200 400 600 8000

Displacement (mm)


0 005 01 015 02 025 03 035 04

Rotation (rad)


700

600

500

400

300

200

100

Tem

pera

ture

(degC

)



800

700

600

500

400

300

200

100

0

Tem

pera

ture

(degC

)

800

1209060300

Component fracture

J1

C1

J1

C1

(c) (d)

(a)

(b)



Deflection

Stab

le r

egio

n


DynamicCritical








Load jack

Reaction frame

Electrical furnace

Macalloy bars

Testedconnection

Reaction frame

Support beam

αFurnacebar Link

bar Jackbar

CameraCamera


0

40

80

120

160

200

240

280


For

ce (

kN)


55deg Load angle


0

20

40

60

80

100

120

140


For

ce (

kN)


tp = 15 mm

550degC


0

50

100

150

200

250

300

350


For

ce (

kN)

2 Bolt rows

3 Bolt rows20degC

550degC








Simple Connections






0

50

100

150

200

250

0 3 6 9 12 15 18 21Rotation (deg)

Tot

al fo

rce

(kN

)

Test


450degC

550degC

650degC


(a)

(b)


0

50

100

150

200

250

300

0 4 8 12 16 20

Rotation (deg)

0 4 8 12 16 20

Rotation (deg)


20degC0

0 3 6 9 12 15

For

ce (

kN)

20

40

60

80

100

120

140

160

180

450degC

For

ce (

kN)

00

100

10

20

30

40

50

60

70

80

90

5 10 15

550degC

For

ce (

kN)





0

5

10

15

20

25

30

35

40

45

650degC















1 111

(a) (b)

of beam

300

300

100

320

100

20 20 7055 20 20325 70

400

320

CL of beamCL





Endplate in bending

Bolt in tension

M


160

140 CIDECT

AISC

Bolt pullout model

Connection element

Test

120

100

80

60

40

20

00 5 10

Rotation (deg)

Forc

e (k

N)

15 20 25


Conclusion






















References





















Acknowledgements










































Option 2

Option 1






1000200 400 600 8000

Displacement (mm)


0 005 01 015 02 025 03 035 04

Rotation (rad)


700

600

500

400

300

200

100

Tem

pera

ture

(degC

)



800

700

600

500

400

300

200

100

0

Tem

pera

ture

(degC

)

800

1209060300

Component fracture

J1

C1

J1

C1

(c) (d)

(a)

(b)



Deflection

Stab

le r

egio

n


DynamicCritical








Load jack

Reaction frame

Electrical furnace

Macalloy bars

Testedconnection

Reaction frame

Support beam

αFurnacebar Link

bar Jackbar

CameraCamera


0

40

80

120

160

200

240

280


For

ce (

kN)


55deg Load angle


0

20

40

60

80

100

120

140


For

ce (

kN)


tp = 15 mm

550degC


0

50

100

150

200

250

300

350


For

ce (

kN)

2 Bolt rows

3 Bolt rows20degC

550degC








Simple Connections






0

50

100

150

200

250

0 3 6 9 12 15 18 21Rotation (deg)

Tot

al fo

rce

(kN

)

Test


450degC

550degC

650degC


(a)

(b)


0

50

100

150

200

250

300

0 4 8 12 16 20

Rotation (deg)

0 4 8 12 16 20

Rotation (deg)


20degC0

0 3 6 9 12 15

For

ce (

kN)

20

40

60

80

100

120

140

160

180

450degC

For

ce (

kN)

00

100

10

20

30

40

50

60

70

80

90

5 10 15

550degC

For

ce (

kN)





0

5

10

15

20

25

30

35

40

45

650degC















1 111

(a) (b)

of beam

300

300

100

320

100

20 20 7055 20 20325 70

400

320

CL of beamCL





Endplate in bending

Bolt in tension

M


160

140 CIDECT

AISC

Bolt pullout model

Connection element

Test

120

100

80

60

40

20

00 5 10

Rotation (deg)

Forc

e (k

N)

15 20 25


Conclusion






















References





















Acknowledgements










































Option 2

Option 1



Load jack

Reaction frame

Electrical furnace

Macalloy bars

Testedconnection

Reaction frame

Support beam

αFurnacebar Link

bar Jackbar

CameraCamera


0

40

80

120

160

200

240

280


For

ce (

kN)


55deg Load angle


0

20

40

60

80

100

120

140


For

ce (

kN)


tp = 15 mm

550degC


0

50

100

150

200

250

300

350


For

ce (

kN)

2 Bolt rows

3 Bolt rows20degC

550degC








Simple Connections






0

50

100

150

200

250

0 3 6 9 12 15 18 21Rotation (deg)

Tot

al fo

rce

(kN

)

Test


450degC

550degC

650degC


(a)

(b)


0

50

100

150

200

250

300

0 4 8 12 16 20

Rotation (deg)

0 4 8 12 16 20

Rotation (deg)


20degC0

0 3 6 9 12 15

For

ce (

kN)

20

40

60

80

100

120

140

160

180

450degC

For

ce (

kN)

00

100

10

20

30

40

50

60

70

80

90

5 10 15

550degC

For

ce (

kN)





0

5

10

15

20

25

30

35

40

45

650degC















1 111

(a) (b)

of beam

300

300

100

320

100

20 20 7055 20 20325 70

400

320

CL of beamCL





Endplate in bending

Bolt in tension

M


160

140 CIDECT

AISC

Bolt pullout model

Connection element

Test

120

100

80

60

40

20

00 5 10

Rotation (deg)

Forc

e (k

N)

15 20 25


Conclusion






















References





















Acknowledgements










































Option 2

Option 1



0

50

100

150

200

250

0 3 6 9 12 15 18 21Rotation (deg)

Tot

al fo

rce

(kN

)

Test


450degC

550degC

650degC


(a)

(b)


0

50

100

150

200

250

300

0 4 8 12 16 20

Rotation (deg)

0 4 8 12 16 20

Rotation (deg)


20degC0

0 3 6 9 12 15

For

ce (

kN)

20

40

60

80

100

120

140

160

180

450degC

For

ce (

kN)

00

100

10

20

30

40

50

60

70

80

90

5 10 15

550degC

For

ce (

kN)





0

5

10

15

20

25

30

35

40

45

650degC















1 111

(a) (b)

of beam

300

300

100

320

100

20 20 7055 20 20325 70

400

320

CL of beamCL





Endplate in bending

Bolt in tension

M


160

140 CIDECT

AISC

Bolt pullout model

Connection element

Test

120

100

80

60

40

20

00 5 10

Rotation (deg)

Forc

e (k

N)

15 20 25


Conclusion






















References





















Acknowledgements










































Option 2

Option 1








1 111

(a) (b)

of beam

300

300

100

320

100

20 20 7055 20 20325 70

400

320

CL of beamCL





Endplate in bending

Bolt in tension

M


160

140 CIDECT

AISC

Bolt pullout model

Connection element

Test

120

100

80

60

40

20

00 5 10

Rotation (deg)

Forc

e (k

N)

15 20 25


Conclusion






















References





















Acknowledgements










































Option 2

Option 1


















References





















Acknowledgements










































Option 2

Option 1









































Option 2

Option 1

Documents

Structural Fire Engineering: Introduction with max DOI.pdf · Structural fire safety is therefore a key consideration in the design of buildings and is attracting worldwide attention