Luận án tiến sĩ: Behaviour of Steel Concrete Composite Beams Made of Ultra High Performance Concrete

The concrete mix contained either 1% fibres or 0.5% by vol-ume of straight steel fibres with concrete strength of approximately 150 MPa.The experimental study demonstrates that the use of

List of Symbols

Greek characters σ c stress of concrete δ uk characteristic value of slip capacity η degree of shear connection κ curvature φ diameter of concrete dowel

Latin lower case letters b o bottom width of shear surface in dowel area d depth of shearing cone n dw numer of dowel in the Push-Out specimen h sc height of steel rib t sc thickness of steel rib q u shear capacity per perfobond

P dw shearing capacity of plain concrete dowel

P r contribution of rebar in dowel to capacity of PSC

P fr contribution of rebar in front cover to capacity of PSC

P a contribution of steel rib to capacity of PSC

A Cross-sectional area of the effective composite section neglecting concrete in tension

A a cross-sectional area of the structural steel section

A b cross-sectional area of bottom transverse reinforcement

A bh cross-sectional area of bottom transverse reinforcement in a haunch

A c cross-sectional area of concrete

A cc cross-sectional area of concrete shear per connector

A fc cross-sectional area of the compression flange

A r area of embedded reinforcement in concrete dowel

A rf amount area of reinforcement in front cover

P u ultimate resistance of Push-Out specimen

P u , test ultimate resistance of Push-out specimen from test

P u , pred predicted ultimate resistance of Push-out specimen

P Rk ,1 characteristic value of the shear resistance of a single connector

P Rk characteristic value of the shear resistance of Push-Out specimen

Mechanical Properties f c Cylinder compressive strength f c , cube Cube compressive strength (150 mm) f ck Characteristic value of the cylinder compressive strength of concrete f ct Tensile strength of concrete f c ,28 d compressive strength of concrete at 28 days f y Nominal value of the yield strength of structural steel f y , r yield strength of reinforcement

E a elastic modulus of structural steel

G f Fracture Energy l ch Characteristic length ν Possion’s ratio γ v Partial factor for design shear resistance of a shear conector

1.1 Karl-Heine footbridge in Leipzig-Germany: concrete filled steel tube structures, afterKoenig[56] (left), and the composite floor of a residential building in London[26] (right) 1

1.2 Basic mechanism of composite action 2

1.3 Perfobond shear connection in composite beam 4

2.1 Typical cross sections of composite beams [26] 7

2.2 Typical shear connectors, afterOehlers and Bradford[68] 8

2.3 Stages of composite beam at different load levers[26] 10

2.4 Longitudinal shear force on connectors[26] 10

2.5 Typified VFT-WIB composite section (above) and application in Vigaun bridge project, afterSchmitt et al [94] 14

2.6 Push-Out specimens and test setup, a) general specimen (Oguejiofor and Hosain [83]), b) specimen with profile steel sheet (Kim et al.[55]) 16

2.7 Shear transfer mechanism from concrete slab to steel rib 17

2.8 Various kind of Perfobond Shear connector in composite beam 18

2.9 Push-Out test of the VFT-WIB connector [93] 19

2.10 Discrete and continuous model for shear connector in composite beams 23

2.11 Elasto-Fracture-Plastic based material models for steel and concrete in Finite element modelling of Push-Out test and com- posite beam 23

2.12 Push-Out specimen model of Kraus and Wurzer[57] 25

2.13 Ideallized tress-strain diagrams used in the plastic method, [26, 27] 26 2.14 Plastic analysis of composite section under sagging moment, 1a- neutral axis in concrete slab; 1b-neutral axis at the bottom of com- posite slab; 2a-neutral axis lies within top flange of steel section; 2b- neutral axis in the web 27

2.15 Design method for partial shear connection [47, 48] 27

3.2 Comonents of a typical UHPC 343.3 Relative density vesus w/c ratio, afterRichard and Cheyrezy[90] 36

3.4 Estimation cost of constituent materials for UHPC, (a):UHPC without steel fiber, (b) with 1% steel fiber [58] 37

3.5 Autogeneous shrinkage of UHPC with and without coarse aggre- gates, afterMa et al.[69, 70] 39

3.6 Creep of UHPC with and without coarse aggregates, afterMa and

3.7 Porosity of UHPC with and without heat treated, afterCwirzen[23] 41 3.8 Comparison durability properties of NSC, UHP and UHPC After

3.10 Test setup for stress-strain response under uni-axial compression 44 3.11 Loading procedure for uni-axial compression test 44

3.12 A comparison of stress-stress curves of NSC, HPC and UHPC(left), and Poinsson’s ratio (right) After(Tue et al.)[101] 44

3.13 Relation elastic modulus vesus compressive strength.(Tue et al.[101, 70]) 45

3.14 Comparison influence of grain size and fiber content to bi-axial strength increment, modified fromCurbach and Hampel[22] 47 3.15 Proposal reduction strength under compression-tension load, mod- ified from (Fehling et al.[29]) 47

3.16 Flexural tensile stress-deflection diagram of G7-UHPC, by Tue et al.[108] 48

3.17 Notched beam three points bending test(left) and Wedge splitting test (right) to determine fracture energy of concrete 49

3.18 Characteristic length versus versus compressive strength [32] 50

4.1 Behaviour of headed stud shear connector in NSC, after John- son [47] 53

4.2 Standard Push-Off Test, Setup 1 (a) and Setup 1 (b) 55

4.3 Typical stress-strain curve of structural steel at room temperature, modifiedOutinen et al.[85] 56

4.4 Typical stress-strain curves of Bst500 reinforcement 56

4.5 Material responses of G7-UHPC 1% steel fiber, stress-strain di- agram in compression test (left) and stress-deflection in RILEM beam test(right) 58

4.7 CDW (above line) and ODW (below line) shear connectors, (a & e)-without rebar, (b & f)-rebar in dowel, (c & g)-rebar in front cover, (d & h)-rebar in dowel and front cover 604.8 Push-Out specimen in 4000 kN load frame and controller system 614.9 Instrumentation setup in SPOT Setup 1(left) and Setup 2 (right) 61

4.11 Load-slip diagram of headed studs shear connectors in UHPC 64

4.12 Crack opening in concrete surfaces 64

4.13 Failure process and shanked of HSSH at footing 65

4.14 Basic mechanics of perfobond shear connector (left), stress state in concrete dowel, afterKraus and Wurzer[57](right) 66

4.15 Deformation of the steel ribs after test 66

4.16 Overview behaviour of perfobond shear contectors 66

4.17 Load-Slip behaviour of CDW and ODW (1 % steel fiber) 68

4.18 Influence of fiber content on load-slip behaviour series 8: 0.5% and series 9: 1% vol steel fiber 69

4.19 Crack opening curves of series 8 and 9 69

4.20 Crack pattern of SPOT with UHPC 0.5% (left) and 1% (right) steel fiber 70

4.21 Crack on the concrete surface, without reinforcement in cover (left) and with reinforcement(right) 70

4.22 Effect of transverse reinforcement arrangement on load-slip be- haviour 71

4.23 Influence of reinforcement thought dowel 72

5.1 Sketch layout of Beam B1 and B2 77

5.2 Sketch layout of Beam B3 and B4 77

5.3 Design layout of Beam B5 and B6 78

5.4 Instrumentation for flexural test of composite beams Series 1 80

5.5 Instrumentation for flexural test of composite beams Series 2 80

5.6 Equipment for flexural test of composite beams Series 1-2 81

5.7 Force-deflection curve before and after remove residual strain 82

5.8 Load-deflection behaviour of composite beam B1 and B2 83

5.9 Plastic of steel girder and crushed of concrete slab 83

5.10 Moment curvature relationship of beam B1 and B2 85

5.11 Strain development in concrete slab (left) and steel girder(right) of composite beam B1 and B2 86

5.12 Strain development in cross section of composite beam B1 and B2 86 5.13 Longitudinal slip of beam B1 (left) and B2 (right) 87

5.14 Lateral strain surround hole of perforated strip 88

5.15 Load-deflection behaviour of composite beam B3 and B4 89

5.16 Failure of beam B3 due to collapse of shear connector in right side 90 5.17 Load-strain behaviour of composite beam B3 and B4, concrete slab (left) and steel girder (right) 92

5.18 Load-strain development in cross section beam B3(left) and B4(right) 92

5.19 Diagram Load-longitudinal slip in beam B3 and B4 93

5.20 Load - deflection behaviour diagrams of beam B5 94

5.21 Load - strain response of beam B5 95

5.24 Load - deflection diagrams of beam B6, UHPC G7 0.5 % fiber content 97

5.25 Load - slip behaviour of beam B6 98

5.26 Failure progress of composite beam B6 99

5.27 Load-Strain at middle span section of beam B6 100

5.28 Strain development in middle span section (left) and one third section (right) of beam B6 100

5.29 Stress-strain over slab thickness 101

5.30 Comparison load slip behaviour of shear connector in composite beam and push out test 102

5.31 Comparison load slip behaviour of shear connector in composite beam and push out test 102

5.32 Slip distribution versus degree shear connection 103

5.33 Longitudinal shear force in composite beams 103

6.1 Bilinear Elasto-plastic material model for structural steel 108

6.2 Calculation macro stress scheme in microplane model 111

6.3 Strain component on a micro plane 111

6.5 FE simulation RILEM (left) bending test and uni-axial compres- sion (right) 116

6.6 Typical stress-strain of uni-axial compression test (left) and bend- ing stress-displacement diagram of RILEM three points bending test (right) 116

6.7 Effect of changing elastic modulus to flexural and compression specimens 119

6.15 Stress-displacement and Stress-strain response of G7-UHPC (1% vol steel fiber) with Microplane M4 material model adjusted pa- rameters 123

6.16 Stress-displacement and Stress-strain response of B4Q-UHPC (1% vol steel fiber) with Microplane M4 124

7.1 Geometry of push-out test specimens 128

7.2 Finite Element model of Push-Out specimen 129

7.3 Loading, boundary conditions and constrain DOFs at contact sur- faces between steel and concrete 131

7.4 Comparison load-slip response of experimental and FE analysis for Push-Out series 3 and 4 (open dowel) 134

7.5 Comparison load-slip response of experimental and FE analysis for Push-Out series 6 and 7 (closed dowel) 134

7.6 Local deformation of the series 4 - Open dowel with test setup 2 136 7.7 Local deformation of the series 7 - Closed dowel with test setup 1 136 7.8 Local stress distrubution in the steel rib 137

7.9 Local strain distribution in concrete block 138

7.10 Stress concentration distribution in rebars of Series 4 (ODW) and

7.12 Geometry of composite beam for FE modelling 146

7.13 Finite Element mesh of a composite beam model 147

7.14 Interface between steel and concrete surface 147

7.15 Deformed shape of the beam B1 and FE simulation 150

7.16 Comparison test and modelling results of beam B1 and B2, force - deflection 151

7.17 Comparison test and modelling results of beam B3 and B4, force - deflection 151

7.18 Comparison test and modelling results of beam B1, force-strain 152 7.19 Comparison test and modelling results of beam B2, force-strain 152 7.20 Comparison test and modelling results of beam B3, force-strain 153 7.21 Comparison test and modelling results of beam B4, force-strain 153 7.22 Comparison local slip of beam B1 (left) and B2 (right) 154

7.23 Stress distribution in girder, beam B1 to B4 155

7.24 Stress distribution in steel rib 156

7.25 Longgitudinal stress in steel rib of shear connector, beam B1 and

B.1 Push-Out test setup S1 and S2 170

B.2 Rebars arrangement of Push-Out specimens 171

B.3 Push-Out test reults: Load-Slip and Crack opening, Series1-Headed stud shear connector, specimen-1(a), specimen-2(b),specimen-3(c) 172

B.4 Push-Out test reults: Load-Slip, Series 2-ODW without rebar (left), Series 3-ODW with rebar in core(right) 173 B.5 Push-Out test reults: Load-Slip and Crack opening, Series 4-Open dowel with rebar in core and front cover, specimen-1(a), specimen- 2(b), specimen-3(c) 174 B.6 Push-Out test reults: Load-Slip and Crack opening, Series 5-CDW without Reinforcement, specimen-1(a), specimen-2(b), specimen- 3(c) 175 B.7 Push-Out test reults: Load-Slip and Crack opening, Series 6-CDW with rebar in core, specimen-1(a), specimen-2(b), specimen-3(c) 176 B.8 Push-Out test reults: Load-Slip and Crack opening, Series 7-Open dowel with rebar in core and front cover, specimen-1(a), specimen- 2(b), specimen-3(c) 177 B.9 Push-Out test reults: Load-Slip and Crack opening, Series 8-

CDW with rebar in cover-UHPC 0.5% steel fiber, specimen-1(a), specimen-2(b) 178 B.10.Push-Out test reults: Load-Slip and Crack opening, Series 9-

CDW with rebar in cover-UHPC 1.0% steel fiber, specimen-1(a), specimen-2(b) 179 B.11.Push-Out test reults: Load-Slip and Crack opening, Series 10-11-

CDW with rebar in core and front cover-UHPC 1.0% steel fiber, φ8mm-(a),φ12mm-(b) 180

C.1 Design of the composite beam B1 182 C.2 Design of the composite beam B2 183 C.3 Design of the composite beam B3 184 C.4 Design of the composite beam B4 185 C.5 Design of the composite beam B5 186 C.6 Design of the composite beam B6 187 C.7 Experimental setup of the composite beam B1 188 C.8 Experimental setup of the composite beam B2 189 C.9 Experimental setup of the composite beam B3 190 C.10.Experimental setup of the composite beam B4 191 C.11.Experimental setup of the composite beam B5 and B6 192 C.12.Beam B1, Load-deflection and Load-rotation (a), strain in girder section 1-1 (b) and strain in girder section 2-2 (c) 193 C.13.Beam B1, Load-strain in concrete slab (a), strain in steel rib (b) and slip (c) 194 C.14.Beam B2, Load-deflection and Load-rotation (a), strain in girder section 1-1 (b) and strain in girder section 2-2 (c) 195

C.15.Beam B2, Load-strain in concrete slab (a), strain in steel rib (b) and slip (c) 196 C.16.Beam B3, Load-deflection and Load-rotation (a), strain in girder section 1-1 (b) and strain in girder section 2-2 (c) 197 C.17.Beam B3, Load-strain in concrete slab (a), strain in steel rib (b) and slip (c) 198 C.18.Beam B4, Load-deflection and Load-rotation (a), strain in girder section 1-1 (b) and strain in girder section 2-2 (c) 199 C.19.Beam B4, Load-strain in concrete slab (a), strain in steel rib (b) and slip (c) 200 C.20.Beam B5, Load-deflection (left), strain in girder and concrete slab at section 1-1 (right) 201 C.21.Beam B6, Load-deflection and Load-rotation (a), strain in girder and concrete slab section 1-1 (b) 201 C.22.Beam B6, strain in girder and concrete slab section 2-2 (a), Load- longitudinal slip along left and right side of the beam (b) 202

D.1 Structure of the program 203D.2 Flow chart of calibration model parameter of microplane M4 204D.3 Main screen of the program 205D.4 Result extraction 205D.5 Quick plot experiment results 206D.6 Atena datafile editor 206

List of Tables

Introduction

The term Composite Construction is normally understood within the context of buildings and other civil engineering structures, to imply the use of Steel and Concrete combine together as a unified component The aim is to archive a higher level of performance than would be have been the case had the two materials functioned separately Steel and concrete can be used in mixed structural sys- tems, for example concrete cores encircled by steel tubes, concrete slab ”glued”’ with steel girder via shear connection in order to form composite beam which most widely used in practical construction Moreover, composite columns offer many advantages over bare steel or reinforced columns, particularly in reducing column cross-sectional area Another important consideration is fire resistance.

Figure 1.1 shows Karl-Heine pedestrian bridge in Leipzig (Koenig[56]), and the composite floor of a residential building in London [26] They are the typical illustration of using hybrid structures in construction.

Figure 1.1.: Karl-Heine footbridge in Leipzig-Germany: concrete filled steel tube structures, after Koenig [56] (left), and the composite floor of a residential building in London[26] (right)

The basic mechanics of composite action is best illustrated by analysis a com- posite beam under bending load which demonstrated in Fig 1.2 In the case of non-composite (a), the concrete slab is not connected to the steel section and therefore behaves independently As it is generally very weak in longitudinal bending it deforms to the curvature of the steel section and has its own neutral axis The bottom surface of the concrete slab is free to slide over the top flange of the steel section and considerable slip occurs between the two The bending resistance of the slab is often so small that it is ignored.

Alternatively, if the concrete slab is connected to the steel section (b), both act together in carrying the service load Slip between the slab and steel section is now prevented and the connection resists a longitudinal shear force Conse- quently, the load bearing capacity of the second beam (b) is few times greater than the first beam (a).

Figure 1.2.: Basic mechanism of composite action

The characteristic of the steel-concrete composite is exhibited by resistance of each contributed material portion and the resistance of shear connection When the connection cannot resist all of the forces applied then considered as partial connection, otherwise full shear connection.

Most frequently, composite beam is designed to carry bending load Regarding the stress and strain distribution of composite section as shown in Fig 1.2b,

1.2 Context and motivation 3 the neutral axis dose not often fall at the interface Good design will attempt locate this axis close to this position Thus whole concrete slab is subjected to compressive force, whereas steel girder to be concerned tension force In prac- tical constructions, the composite beam is often made of either normal strength concrete (in short NSC) or high strength concrete (in short HSC) for slab and high strength steel for girder.

Recent development of concrete technology resulting a new type of concrete with many advanced properties, it is called in common name Ultra High Performance Concrete (in short UHPC) The key benefits of UHPC are considered in applica- tion point of view as follows: very high in compressive strength and tensile strength which are ideal to carry compression load in the composite beams. addition steel fiber will enhanced ductility behaviour reduce total weight of structural member with high flow ability properties, concrete can be complete fulled for com- plex geometry members. extraordinary durability compare to conventional concrete, reduce maintain cost during service time. most disadvantage of UHPC is highly cost at the moment, it may be de- creasing in the near future when increasing amount of applications The detail characteristic of UHPC will be mentioned in the chapter 3.

In the structural member behaviour outlook, with NSC the resistance of concrete slab is often less than steel girder, the neutral line lie in the web.

By substituting UHPC to NSC/HSC, the resistance of concrete materials could be reached resistance capacity of steel easily Consequently obtaining optimal load caring of each contribute material The replacement is not only increase the stiffness and overall ultimate strength but also reduce cross section of the composite beams.

Fig 1.3 illustrates the idea using perforated steel rib as continuous shear connec- tion in the composite beam This type of shear connector was first introduced byLeonhardt[62] Perforation strip are welded on top flange of steel girder or cut directly from web At construction phase, UHPC will be flowed through perfo- rated hole the dowels formed Under loading, interaction is developed by concrete engaging with perforations strip, the working mechanism of shear connector can be illustrated similar to the action of a dowel In principle, this method brings to many advantages in practical construction, while load transfer performance is still guaranteed.

Figure 1.3.: Perfobond shear connection in composite beam

It is well known that, at interaction area between perforated strip and UHPC dowel, the behaviour is combination of tension-shear and compression The UHPC with very high compressive strength but less ductility must be treated to satisfy characteristic ductility requirement of shear connector in composite beam The application of this device for shear connection incorporating steel girder still requires further verification.

Due to the high cost of UHPC material and testing, the experimental study is unable to cover all range of interested problems Consequently, numerical simulation play an important role in this works However, the behaviour of UHPC is different with conventional concrete, therefore suitable material model is required to illustrate mechanism of beam as well concerned problems.

The present study aims to investigate performance and structural behaviour of steel-concrete composite beam made of UHPC under bending, and it also provide

1.4 Scope of work 5 a better knowledge of perfobond shear connector response in Push-Out test and conjugate with steel girder More precisely, the following points are explored:

Characteristic of UHPC would be better known and understood, especially focus on basic mechanical properties.

A better knowledge on response of the perfobond shear connectors in UHPC, appropriate choice of shear connector for UHPC composite member would be achieved.

Experimental investigation of UHPC composite beam subjected flexural load, which provides structural behaviour of member under serviceability and ultimate limit state, in order to answer the following questions:

- Is it possible to build composite ”UHPC-Steel” elements with monolithic behaviour; and how can the advantageous UHPC properties be exploited in such composite elements?

- What do resistance and failure modes of ”UHPC-Steel” elements would be shown under bending?

- How do local deformations, stresses and cracking evolve in the composite members under monotonic load?

Nonlinear finite element models must be evaluated and developed in order to predict the structural behaviour of shear connectors and ”UHPC-Steel” composite beams The simulation should be explored following aspects:

- Are existing material models appropriate to simulate behaviour of UHPC?

- How to construct suitable structural models for shear connector and com- posite beams?

- What do local behaviour would be shown?

- How to improve performance of the ”UHPC-Steel” composite beam?

On the basis of the results, a design model and guidelines are developed for practical application of UHPC composite members.

This work is part of priority research program SPP 1182: ”Sustainable Building with Ultra High Performance Concrete”, which collaborate by numerous of uni- versities in Germany The concrete material and design of composite were prior planned, and oriented to the trend of this project The flexural behaviour of sin- gle span composite beams were limited to sagging moment only The continuous beam with hogging moment (negative moment) at support is not considered in this work.

The thesis consists of eight chapters Chapter one is the outline introduction to innovation context of development of UHPC and its application into hybrid steel-concrete structures The main aspect and objective of this research work was also mentioned.

Chapter two presents relevant literature review of the behaviour of steel-concrete composite beams made of UHPC The content includes material properties as- pect, load transfer mechanism in the beam, as well as experiment and modelling of composite beams.

In Chapter three, the state of the art of UHPC are brief introduced, properties of UHPC are characterized and main properties which influence on behaviour of structures under loading service are to be discussed in details.

Consideration aspects of steel-concrete composite beamscomposite beams

The most important and most frequently encountered combination of construc- tion materials is that of steel and concrete, with applications in multi-storey buildings and constructions, as well as in bridges These materials can be used in mixed structural systems, for example concrete slab glued with steel girder, as well as in composite structures where members consisting of steel and concrete act together Steel and concrete have the same expansion coefficient, and each materials is strong in either compression or tension Concrete also provides cor- rosion protection and thermal insulation to the steel at elevated temperatures and additionally can restrain slender steel sections from local or lateral-torsional buckling These essentially different materials are completely compatible and complementary to each other.

Composite beams, subjected mainly to bending, consist of a steel section acting compositely with one (or two) flanges of reinforced concrete The two materials are interconnected by means of mechanical shear connectors For single span beams, sagging bending moments, due to applied vertical loads, cause tensile forces in the steel section and compression in the concrete deck thereby allowing optimum use of each material Fig 2.1 and Fig.2.2 show several composite beam cross-sections and shear connectors respectively, which are widely used in practical construction.

Figure 2.1.: Typical cross sections of composite beams [26]

Figure 2.2.: Typical shear connectors, after Oehlers and Bradford [68]

The shear connectors in composite beams are used to develop the composite action between steel girder and concrete They are provided mainly to resist longitudinal shear force, therefore must meet a various requirements, such as [26]: transfer direct shear at their base. create a tensile link into the concrete. economic to manufacture and welding.

The most common type of mechanical shear connector is the headed stud shown in Fig 2.2a It can be welded to the upper flange either directly in the factory or through thin galvanised steel sheeting on site The Behaviour and ultimate strength of connectors can be examined by Push-Out test according to available standards such as EuroCode4 [27] For the design of headed stud, the fol- lowing aspects are considered; shear strength of stud shank, bearing strength of concrete, additional contribution of chemical bonding and friction In spite of its wide application, the headed stud has many deficiencies such as a slip Behaviour between stud and concrete, and fatigue failure at welding zone [26, 80, 47, 55]

Recently, a very high strength cement based composite called Ultra High Perfor- mance Concrete (UHPC) has been developed It provides many enhancements in properties compared to conventional and high strength concrete (HSC) In

2.2 Single span composite beams under sagging moment 9 the composite beams, the replacement of normal strength concrete (NSC) with UHPC lead to an improvement in the load carrying in the compression zone.

Generally, a significant increase in load bearing capacity and stiffness of the beam is achieved, resulting in saving dead load, reducing construction depth as well as construction time However, as reported in Johnson[47], Hegger et al., Tue et al [105] the headed stud shear connector is not appropriate in the HSC/UHPC slab due to restrict deformation surrounding stud area The combination of perfobond shear connector in UHPC will be optimized in both term of material and structural system.

This chapter aims to review researches relevant to the Behaviour of composite beams under bending load, which focuses to composite beam/slab with perfobond shear connector Different aspects of the problem are discussed such as the basic Behaviour of composite beams, innovation of concrete technology, mechanical shear connection The numerical modelling of the structural composite beams and the currently available design procedures will be also mentioned.

2.2 Single span composite beams under sagging moment

The way in which a composite beam behaves under the action of low load, medium and the final failure load can be briefly described in stages as follows [26]:

Under very low loads the steel and concrete behaves in an approximately linear way The connection between the two materials carries very low shear stresses and it is unlikely that appreciable longitudinal slip will occur The beam deforms so that the strain distribution at mid span is linear, as in Fig 2.3a, and the resulting stress is also linear.

It can be seen from the strain diagram that, if the slab is thick enough then the neutral axis lies within the concrete As a result some of the concrete is in tension If the slab was thin, it is possible that the neutral axis would be in the steel and then the area of steel above the axis would be in compression This stage corresponds to the service load situation in the sagging moment region of most practical composite beams.

Figure 2.3.: Stages of composite beam at different load levers[26]

Figure 2.4.: Longitudinal shear force on connectors[26]

In this stage applied load was increased, thus caused rise to deformation in the shear connection This deformation is known asslipand contributes to the overall

2.2 Single span composite beams under sagging moment 11 deformation of the beam Fig 2.3b shows the influence of slip on the strain and stress distribution This stage corresponds to the service load stage that composite beams class has been designed as partially shear connection However, for many composite beams slip is very small and may be neglected.

The steel girder achieves yield limit strain first, plasticity develops and then al- most part of steel section becomes plastic It occurs as similar fashion in concrete slab Stress block of whole section changes from triangular to shape shown in Fig 2.3c that is very difficult to express in mathematical form In ultimate limit state (ULS) it is assumed to be a rectangular block.

If longitudinal shear resistance is big enough the slip can be neglected The strain in concrete slab could lead to over stress, then it is potentially possible that explosive brittle failure of the slab would occur However, in most practical case this situation could ever arise due to the deformation of shear connectors.

The response of shear connector in load stage is illustrated as follows:

As the load increases the shear strain, the longitudinal shear force between the concrete slab and steel girder increases in proportion For single span composite beam under uniformly load, it is assumed to deform in an elastic manner and the longitudinal shear force between slab and steel section can be expressed as

T =VS/I [96] Hence longitudinal shear force is directly proportional to the vertical shear force, thus the force on the end connectors is the greatest For low loads the force acting on a connector produces elastic deformation The slip between the slab and the steel section will be greatest at the end of the beam.

The longitudinal shear and deformation of a typical composite beam, at this stage of loading, are shown in Fig 2.4a.

Characterization material properties of UHPC

3.1 Development of UHPC-A Historical perspective

The development of high strength concrete began in 1970s, when the first time the compressive strength of the concrete used in the columns of some high rise building was higher than that of concrete usually used in construction The concrete was made using the same technology as that for normal strength concrete expect that the materials were carefully selected and controlled [4] The new conceptHigh Strength Concrete-HSC was called for this concrete.

With the development of superplasticizers and the usage of pozzolannic admix- trure such as silica fume, it is possible to produce concrete with compressive strength more than 150 MPa [4] Moreover, the concrete has also improved characteristics such as higher flowability, elastic modulus, flexural strength, low permeability and better durability over NSC.

The expression High Strength Concrete can no longer adequately describe the overall improvement in the properties Therefore, the new expressionHigh Per- formance Concrete - HPC became more widely used early 1990s [78, 4, 81, 1].

ACI 363 committee defined HPC as follows:

HPC is concrete meeting special combinations of performance and uniformity requirements that can not always be archived routinely using only conventional constituent and normal mixing, placing and curing practices These requirements may involve enhancements of the following [43]: easy of placement and compaction without segregation long term mechanical properties early age strength toughness volume stability long life in sever environments

The development of material technology in the early 2000s not only enhance quality but also reduce significantly their cost In fact, HPC was used widely in many applications Up to now, in the normal curing condition a compressive strength can be reach over 200 MPa However, in this case, the concrete is very brittle Consequently, the addition of fiber is necessary to improve the ductility Since, the the new concept Ultra High Performance Concrete-UHPC began widely used [56, 111, 4], Fig 3.1 summarize the historical development of concrete.

Figure 3.1.: Historical development of UHPC

Several types of UHPC have been developed in different countries by different manufactures or research institutions Some product lines have been marketed and became commercialize There are few major types of UHPC those namely Ceracem/BSI (by Sika)[77], compact reinforced composites (CRC) by CRC Tech- nology [52], multi-scale cement composite (MSCC) by Laboratoire Central des Ponts et Chausees (France) [91], and reactive powder concrete (RPC) by Lafarge as known with commercial name DUCTAL [3] This count is by no means a complete overview of all mixtures, as more mixtures are being developed and entering the market from different laboratories and universities.

All the above described mixtures were designed with the main aim to reach a high compressive strength, while the improvement of the tensile and flexural tensile strength was of secondary interest Another, different group of fibre reinforced concretes has also been developed where instead of the compressive strength,

3.2 Constituent materials of Ultra High Performance Concrete 33 the focus was set on improving the tensile load bearing capacity, and especially the tensile deformation capacity These ductile concretes are often called High Performance Fibre Reinforced Cementitious Composites-HPFRCC or commonly UHPC [60, 63].

In Germany, the research program on UHPC was carried out early ten years ago[56] Especially, the priority research project SPP 1182 -Sustainable Building with Ultra High Performance Concretehas been performed with the collaboration of many research institutions This work is also a part of this project.

3.2 Constituent materials of Ultra High Performance Concrete

Several authors have been identified the basic principles to produce UHPC, which can be summarized as follows [99]: enhancement of homogeneity by elimination of coarse aggregate. enhancement of the packing density by optimization of the granular mixture through a wide distribution of powder size classes. improvement of the properties of the matrix by the addition of pozzolanic admixture, such as silica fume. improvement of the matrix properties by reducing water/binder ratio. enhancement of the microstructure by post-set heat-treatment, and enhancement of ductility by addition of micro steel fibers.

The application of the fist five principles lead to very high compressive strength, however without any improvement in ductility UHPC could be cured with high temperature and pressure condition after setting High pressure treatment in- creases density by reducing entrapped air, removing excess water and acceler- ating chemical shrinkage Heat treatment accelerates the cement hydration and puzzolanic reaction as well as modifies micro structures of the hydrates [36, 88].

The addition of the steel fibers that noted in the last principle helps to improve both the tensile strength and ductility, whereas polymer and carbon fiber enhancefire resistance The UHPC in this work contains steel fiber but without using any special treatment.

A typical UHPC consists of cement, silica fume, coarse aggregate, sand, crushed quatz, superplasticizer, fiber, crushed quartz, fibers, superplasticizer, and water as well [90].

Figure 3.2.: Comonents of a typical UHPC

Table 3.1.: Diameter range of granular class for UHPC, after Richard and Cheyrezy [90]

Components Mean diameter Typical diameter range

Figure 3.2 shows typical components for make an UHPC and their sizes are presented in table 3.1 The role of the each constituent is briefly summarized as follows:

Cement: Usually ordinary Portland cement type I (CEM I 42.5R/52.5R) or

Portland cement with high sulphate resistance (CEM I 42.5R HS/52.5R HS) can be used to produce UHPC The cement used should be low to medium fineness and not rich inC 3 Acontent Thus, reducing water need ettringite formation and heat of hydration [37] Low shrinkage cements may also be preferred since the high cement content of UHPC can make it more susceptible to high shrinkage[99].

3.2 Constituent materials of Ultra High Performance Concrete 35

Sand: It plays the role of reducing the matrix volume fraction under condition of enough flowability Its strength is higher than the matrix and provides good paste-aggregate interfacing bonding A variety of sand is usually used, however, it is not chemically active in the cement hydration reaction at room temperature.

The mean particle size is often smaller than 1mm It is noted that, the grain size of the silica fume, cement and sand must have to be optimized in oder to get high compact, dense matrix and low permeability.

Crushed Quartz: In fact, not all of cement in the concrete mix is hydrated, some of which can be replaced by crushed quartz powder Ma and Schnei- der[72] pointed out that, up to 30 percent of cement can be replaced by quartz power without reduction in compressive strength Besides that, it also improves flowability of fresh UHPC The improvement of flowability may due to the filling effect, since the crushed quartz particles are smaller than cement particles.

Silica fume: Silica fume is composed of very small of glassy silica particle which are perfectly spherical, whose mean particles is in the range of 0.1 to 1.0 μm.

Silica fume has three main roles in UHPC [99]:

Experimental study for perfobond shear connector in UHPCconnector in UHPC

The behaviour of steel-concrete composite beams in bending is achieved by means of the shear connectors, which play an important role in resisting the longitudinal slip and the separation of concrete slab and steel girder The types and quan- tities of shear connectors depend on the shear force resulting from the bending moment and the vertical loading Conventional shear connection in composite construction is often designed as headed stud, lying stud etc., and normal strength concrete is usually used for slab [47, 80, 27] In normal strength concrete, headed stud shear connector (HSSH) results in high ductile response as shown in Fig.

4.1 If the strength of the concrete surrounding stud are very high, then the defor- mation of the stud is restrained and shear connector can be shanked at base The failure mode is brittle, the ductility is insufficient as required of several design codes.

Figure 4.1.: Behaviour of headed stud shear connector in NSC, after Johnson [47]

The perfobond shear connector (PFSH) was first introduced byLeonhardt[62] in Germany With this kind of shear connector, the interaction is developed by concrete dowel engaging with the perforated steel strip In fabrication the steel strip is cut and attached by welding to steel girder The main advantages of the perfobond shear connection are listed as follows: the carry load can be transferred continuously between concrete slab and steel girder. the same material can be used for shear connector and steel beam, it does not require a higher steel grade for shear connector and special equipment for welding. with symmetric dowel profile two shear connector strips could be receive with only one cutting line and there is no material wasted If the cut is carried out in the web of a steel I-girder, two composite beams without an upper flange can be produced [41, 51, 105] the reduction of total cost by less labor work and faster in fabrication

Since the first time appear to now, perfobond shear connectors have been good alternative solutions for conventional headed stud shear connectors Practical experiences and laboratory studies pointed out that, the strength of steel and concrete, the thickness of steel rib, the profile of dowel, the embedding rebar inside dowel as well as reinforcement in front layer etc are important criteria for the load bearing capacity of the perfobond shear connectors.

The Push-Out tests, presented and discussed in the following, aim to investigate the behaviour of the perfobond as well as headed stud shear connectors in UHPC, which are applied in composite beams Its objective was to identify the appli- cability whether brittle shear connection behaviour could occur and to provide possible reinforcing solutions which ensure sufficiently ductile behaviour During the testing, two types of dowel profile, reinforcing arrangement and steel fiber content that control the concrete-related failure modes were investigated Due to limited condition, the experimental study could not cover all interesting aspects, therefore additional modelling work need be done.

Based on these findings, the test data are used to validate the numerical model and the preliminary suggestions for design shear connection are established.

The standard Push-Out test (here after SPOT) was carried out in order to inves- tigate the behaviour and characteristic parameters of the shear connectors In

4.2 Experimental programs and specimens 55 fact, they were planned to use in the composite beams The testing procedure and the evaluation results of the test were performed according to the guideline of EC4-Appendix B [27] Various test series were prepared for both type of shear connectors: headed stud and perfobond.

Figure 4.2.: Standard Push-Off Test, Setup 1 (a) and Setup 1 (b)

Fig.4.2 depicts the detailed components of a specimen There are three main parts include a thick steel plate, a perforated steel strip and a concrete block.

The steel plate of 200mm width, 350mm height and 20/30mm thickness is pre- sented for the flange (I section) or web (in T section) in steel girder Its stiffness must be strong enough to ensure the transfer of shear force from flange/web to perforated steel strip and concrete dowels The perforated strip with dimensions65/75mm×310mm and 10mm thickness were considered as steel rib of the per- fobond connector As depicted in the figure, each steel rib has two holes of 45mm diameter through which UHPC will flow to form the concrete dowels The profile of dowel was designed with two variants namely closed dowel(CDW) and open dowel (ODW), as illustrated are described in Fig 4.7 The concrete block was300mm wide, 350mm high and 80mm thick that acts the concrete slab in com- posite beams And it includes two dowels which are used to against shear force from the steel rib.

Generally, all specimens are symmetric with identical dimensions of concrete bock and steel parts as well as thickness of steel rib In the specimens using headed stud shear connector, two steel ribs were replaced by eightφ16mm×60mmheaded studs, which welded into the flange or web directly.

Figure 4.3.: Typical stress-strain curve of structural steel at room temperature, modified Out- inen et al [85]

Dia 12mm Dia 10mm Dia 8mm Dia 6mm

Figure 4.4.: Typical stress-strain curves of Bst500 reinforcement

In this study, structural steel grade S355 was used for both Push-Out (PO) specimens and composite beams The mechanical properties of this steel were determined from tensile test However, there were no test for steel plates, all test

4.2 Experimental programs and specimens 57 data were archived from research work carried out byOutinen et al[85] and Byfield et al[13] The values of yield strength, elastic modulus and ultimate strength were evaluated at 380 MPa, 506 MPa and 202.6 GPa, respectively The typical stress strain curve are shown in Fig 4.3 The details of the mechanical properties are given in table 4.1.

Table 4.1.: Mechanical properties of steel grade S355 and reinforcing bar Bst 500

Yield strength f sy (MPa) 386 520.00 Ultimate strength - f su (MPa) 506 600.00

Elastic Modulus - E s (MPa) 202,590 210,000 Hardening Modulus - E sh (MPa) 2,235 - Elongation after fracture (%) 24 -

Strain hardening ε sh (%) 1.50 - Ultimate limit strain ε su (%) 4.00 -

Bst500 grade reinforcement was used for all specimens In order to obtain the essential characteristics, tension test were carried out for rebar with diameter of φ6mm,φ8mm,φ10mm and 12 as well Average values of yield, ultimate strength and limit yield strain of reinforcing bar fromφ8mm toφ12mm are 520 Mp, 600

Mpa and 0.22% respectively The typical stress-strain curves are plotted in Fig.

4.4 and the main mechanical properties are also listed in Tab 4.1.

In the experimental framework of composite beams and Push-Out test, the UPHC G7 mix proportion was used for various test series The details of material com- position was given in previous chapter The steel fiber content was specified with 0.5% (G7-150-0.5%) and 1% (G7-150-1.0%) in order to investigate the influence of tensile toughness of concrete on the specimen behaviour The typical material response curve of G7-UHPC in uni-axial compression and three points bending stress states are shown in Fig 4.5 The basic properties of G7-UHPC are given in table 4.2.

Table 4.2.: Material properties of UHPC

Compressive Elastic Flexural Elastic Limit Concrete strength modulus strength strain strain

∗ UHPC B4Q mixture is used in bending testing of the composite beams B1 to B4

Figure 4.5.: Material responses of G7-UHPC 1% steel fiber, stress-strain diagram in compression test (left) and stress-deflection in RILEM beam test(right)

Figure 4.6.: Casting Push-Out specimens

The specimens of each individual test series were prepared and cast in the vertical direction from the same batch of concrete Numerous of concrete cylinders of φ100mm×200mm were also cast and stored alongside the specimen and tested at

4.2 Experimental programs and specimens 59 regular intervals.

At the stage of producing PO specimens, the gaps at bottom of steel ribs with dimension of 20×20mm×70mm were early created by two foam blocks The aim is to ensure that the steel flange/web and concrete block are properly relative slip in the push out test Further, resistant force will occur only at interface areas between concrete dowel and steel rib Before test these holes were checked again.

Fig 4.6 depicts the form work, the rebar arrangement and the specimen ready for test.

4.2.2 Arrangement for Push-Out series

Parameters investigated in the experiment program include the profile of dowel, the embedded rebar in UHPC dowel, the transverse reinforcement in cover layer, as well as the content of steel fiber in concrete Besides, the headed stud shear connector was also examined in order to compare the conventional and the novel shear connection in UHPC.

Table 4.3.: Parameter for Push-Out test program

Series Concrete Setup NOS ∗ Rebar ∗∗ Description

3 - S2 3 RA ODW with rebar in dowel

4 - S2 3 RAB ODW with rebar in dowel and cover

6 - S1 3 RA CDW with rebar in dowel

7 - S1 3 RAB CDW with rebar in dowel and cover

8 G7-150-0.5% S2 2 RB CDW with rebar in cover

9 G7-150-1.0% S2 2 RB CDW with rebar in cover

10 G7-150-0.5% S2 1 RAB CDW with rebar in dowel and cover

11 G7-150-0.5% S2 1 RAB CDW with φ 12mm rebar in dowel and φ 8mm in cover

Experimental investigation on the structural behaviour of steel-UHPC composite beamsbehaviour of steel-UHPC composite beams

In this chapter will be introduce detail of experimental study of composite beams which were carried out at University of Leipzig (Uni-Leipzig) The composite beam was made of ultra high performance concrete and high strength structural steel Several test series were conducted to obtained overview behaviour as well as to ensure the feasibility of this new structure The relevant such as structural behaviour of composite beams and failure mode, load bearing capacity in ulti- mate and serviceability limit states, load-slip response of shear connection will be mentioned

5.2 Experimental program for composite beams

The experimental study aimed to evaluate the structural response of the Steel- UHPC composite beams under static load Through large scale test some aspects below could be understood: global structural behaviour of the composite beam under static load mode of failure and response on each material part (development of stress- strain) relation between applied load, bending moment and relative slip key parameters governing to global behaviour of beam local strain development in steel rib verify performance of UHPC perfobond shear connectorsThe test data is also to be used on validation the numerical model.

5.2.2 Design and construction of test specimens

The composite beams for testing were designed according toEC4 [27] (section 6), in fact simple plastic method was used The connection between steel and UHPC slab was assumed as full shear connection In the design calculation, all material and load factors were set to unity The beams were divided into two series: series one includes 04 beams (B1 to B4), remaining series has 02 beams (B5 and B6) The steel girder I and T sections were used and UHPC was used for concrete slab Table 5.1 gives a short description of each beams in the test series.

In the case of I steel girders, shear connectors were welded to the top flange by continuous steel strip Otherwise, for the T girder, shear connectors were formed as an extension of the web and cut directly The spacing between two holes are 100 mm or 150 mm depending on the specific purpose of the test, and the hole is 45 mm diameter for all beams.

Table 5.1.: Description of composite beams

Series Beam ID Shear Connector Steel girder section and spacing and concrete slab

Series 1 B1 Open dowel, 59 x 100 mm I shape, 500 x 100 mm - B2 Closed dowel, 59 x 100 mm I shape, 500 x 100 mm - B3 Closed dowel, 39 x 150 mm inversed T, 500 x 100 mm - B4 Open dowel, 39 x 150 mm inversed T, 500 x 100 mm

Series 2 B5 Closed dowel, 79 x 100 mm inversed T, 400 x 100 mm - B6 Open dowel, 79 x 100 mm inversed T, 400 x 100 mm

Table 5.2.: Transverse reinforcement arrangement in concrete slab

Series Beam ID Reinforcement Reinforcement in font layer embedding in dowel

Series 1 B1 φ 8 mm @ 100 mm 8 mm @ 200 mm (twice dowel) - B2 φ 8 mm @ 100 mm 8 mm @ 200 mm (twice dowel) - B3 φ 8 mm @ 100 mm 8 mm @ 150 mm (each dowel) - B4 φ 8 mm @ 100 mm 8 mm @ 150 mm (each dowel)

Series 2 B5 φ 8 mm @ 80 mm no reinforcement - B6 φ 8 mm @ 100 mm 8 mm @ 100 mm (each dowel)

Transverse reinforcing rebar with diameter of 8 mm were placed in two layers at top and bottom of the UHPC slab The top layer had a spacing of 80 mm/100 mm and the bottom layer had spacing of 100 mm/150 mm respective to that of shear connectors The longitudinal reinforcement was arranged with fourφ10

5.2 Experimental program for composite beams 77 mm rebar placed in both sides of the shear connectors Table 5.2 summarizes the arrangement of the transverse reinforcement in each beam.

Figure 5.1.: Sketch layout of Beam B1 and B2

Figure 5.2.: Sketch layout of Beam B3 and B4

S355 structural steel and Bst500 grade reinforcement were utilized to produce all composite beams, whose material properties are identical with steel of Push-Out test Additionly, the B4Q-UHPC mixture was used to made 04 beams of series 1.

And all beams of Series 2 was cast with G7-UHPC mixture All UHPC mixtures contain coarse aggregate (2-5 mm and 5-8 mm) and steel fiber of 0.5% (G7) and 1.0% (B4Q) The primary mechanical properties of both concrete are listed in table 4.2 Further details of material compositions were given in 3.2.4, table 3.2.

The number of shear connector was determined based on the Push-Out tests data Unfortunately, it is not always available due to some out of controlled reasons Thus, in the cases of beams of series 2, the result of Push-Out test came too late Therefore, its is not insufficient information for making right decision during design progress The beam B5 was designed without reinforcement in dowel, which lead to less longitudinal shear resistance However, hence several wrongs good lessons were obtained.

Fig 5.1 depicts the design layout of the composite beams B1 and B2 Both beams were 6.0m in length and 410 mm in height Moreover the pairs of the beam are the same of cross section with I steel girder of 300×310×10×14 mm and concrete slab of 500×100 mm The difference between two beams is only in profile of shear connector, beam B1 and B2 were designed with ODW and CDW, respectively These beams aimed to reach full plastic moment in steel girder.

More full shear connection degrees are also considered to verify load transfer capacity of dowel.

Figure 5.3.: Design layout of Beam B5 and B6

The sketch of beam B3 and B4 are described in Fig 5.2, they have the same length and concrete slab section with previous beams B1 and B2 The spacing between shear connector was 150 mm which is greater than that of beam B1 andB2 (less shear connection degree) Both beams B3 and B4 were designed with T girder and are only different on bottom flange The flange of beam B3 was 320 mm in width that expected to fail in concrete slab or shear connection While

5.2 Experimental program for composite beams 79 the flange width of beam B4 was 200 mm which is expected to be fail by yielding of steel girder and crushing of concrete in compression zone.

The second series include beam B5 and B6 depicted in Fig 5.3 They were designed and built in the second stage of experimental program The test of series 2 had two purposes: to evaluate potential of reducing fiber contents in UHPC and stress in concrete slab only in compression The strain distribution over height of slab would be nearly constant Due to the lack Push-Out test data, beam B5 was made to contain no reinforcement in concrete dowel The reinforcement was arranged only in top cover layer of concrete slab with spacing of 80 mm The influence of transverse reinforcement in cover layer on lateral shear resistance is also analyzed for beam B5.

To build composite beams, the steel girders were fabricated in factory and trans- ported to laboratory while the rebar and concrete work were performed in labo- ratory Numerous cylinders and cubic were cast to test the mechanical properties of concrete, which were cured beside composite beams.

5.2.3 Test set-up and instrumentation

Large scale experiments were arranged according to four points bending test scheme The hinge support was installed at the North end and the South end was placed on roller support Fig 5.4 and 5.5 sketch the general layout of test setup of Series 1 and 2, respectively The test was generally displacement controlled while the speed varied during testing Each specimen was cycled in a similar fashion as the Push-Out test specimens described earlier (Fig 4.10), i.e at least twenty seven (27) times between 10% and a proof load of about 40% of the expected ultimate bending strength After the last cycle completed, the load was applied continuously until the beam failure occurred or until the load dropped to a significant amount below its maximum value.

There are four basic types of instrumentations utilized in the tests Strain gages was used to capture the strain on the steel girders, the shear connectors and the concrete slab Whereas, linear string potentiometers were used to measure deflec- tion along span The relative slip, strain of concrete, rotation angle at support and opening crack on the concrete surface were measured by linear variable dis- placement transducers (LVDT) Load cell was used to measure live loads applied to the beams When test in progress, all measured data were recorded automat- ically by the 48-channels HBM measuring system Details of instrumentation are described in Fig 5.4 and 5.5 The loading equipment system and typical installed transducers are shown in Fig 5.6.

Figure 5.4.: Instrumentation for flexural test of composite beams Series 1

Figure 5.5.: Instrumentation for flexural test of composite beams Series 2

5.3 Analysis of the test results and observations 81

Figure 5.6.: Equipment for flexural test of composite beams Series 1-2

5.3 Analysis of the test results and observations

Material models for Finite Element Modelling

Steel concrete composite beams are made of three material with different charac- teristics, namely concrete, structural steel and reinforcing bars Steel and rebar can be considered as homogeneous materials whose properties are generally well defined Concrete is, on the other hand, heterogeneous material made of many constituents Its mechanical properties scatter more widely and can not be de- fined easily.

Let consider load-slip diagram of push out test and load-deformation of com- posite beam, which were shown in Fig 4.16 and Fig 5.8 respectively It can be easily identified that, the behaviour of these structures are highly nonlinear response It can be roughly divided into three range: un-cracked elastic stage, crack propagation and the plastic (yielding or crushing).

The nonlinear response is caused by two major effects, namely cracking of concrete in tension (such as UHPC dowel region), and yielding of the steel girder/reinforcement or crushing of concrete in compression zone of slab More- over, nonlinearities also arise from interaction of parts of structures, such as bond-slip between reinforcing bar and concrete, aggregate interlock at cracks, dowel action of reinforcing steel crossing a crack and interface between steel and the concrete surfaces, etc The time-dependent effects also contribute to non- linear behaviour Furthermore, the stress-strain relation of concrete is not only nonlinear, but also different in compression and in tension and the mechanical properties are dependent on concrete age at loading and on environmental con- ditions The material properties of steel and concrete are also greatly different [59].

From structural engineering point of view, it is difficult to understand the com- plete mechanics of a structural response of composite beam as well shear connec- tor from experiments alone Within this study, numerical analysis has become increasingly important in obtaining an improved understanding on structural behaviour such as load transfer from concrete slab to steel girder, distribution of longitudinal shear force, stress field in dowel region etc With the advent of high speed computers and special tool for simulation, the nonlinearity of material should be taken into account, in order to describe nonlinear response of struc- tures under external load Some numerical investigation will be conducted with variant of model parameters.

The first part of this chapter will short introduce the material models for struc- tural steel and reinforcement Then next will focus microplane model M4 for concrete The uni-axial compression and bending of three points notched beam will be modeled and analyzed in order to evaluate influence of each parameter on numerical results A procedure for adjusting key parameters of microplane M4 for UHPC was proposed.

6.2 Material models for structural steel and reinforcement

Structural steel is generally a homogeneous material, therefore the specification of single stress-strain relation is usually sufficient to defined the material properties needed in analysis.

In this study, structural steel is modelled as an elastic-plastic material incorpo- rating strain hardening Figure 6.1 shows the stress strain diagram for steel in tension Specifically, the relationship is linearly elastic up to yielding (f sy ), linear hardening occurs up to the ultimate tensile (f su ) stress and the stress remains constant until the tensile failure strain is reached For all practical purposes the steel also exhibits the same way in compression.

Figure 6.1.: Bilinear Elasto-plastic material model for structural steel

For the modelling of reinforcing bar, the elastic-plastic material model for struc- tural steel is used with small modification in yield plateau portion Hwak and

6.3 Microplane M4 material model for concrete 109

Filippou[59],Chen et al.[20] pointed out that, since steel reinforcement have been used in concrete construction as form of rebar or wire, it is not necessary to introduce the complexities of three-dimensional constitutive model for steel.

In this study, the simulation work would be concentrated to Push-Out and com- posite beams tests In fact Push-Out tests requires considering the local be- haviour caused by larger deformation at dowel region Whereas composite beams modelling demands to take into account the global response and local longitudi- nal slips To utilize the computational efficiencies and achieve reasonable results, all most structural models will be discretized by 3D solid element (brick element).

The reinforcement in Push-Out test will be idealized by 3D solid element, whilst in the composite beam the one dimensional stress-strain for reinforcing bar is used.

The deformed reinforcement Bst 500 grade was used for all specimens of compos- ite beams and Push-Out tests Its mechanical properties are given in table 4.1.

According to the design of the beams, most of reinforcing bars were arranged in front surfaces, which lie in compression fiber of UHCP slab Therefore the bond interaction effect between reinforcing steel and surrounding concrete is omitted and perfect bonding is assumed in the analysis.

6.3 Microplane M4 material model for concrete

6.3.1 Aspects of concrete material model

Many constitutive models for describing the mechanical behaviour of concrete are currently in use in the analysis of reinforcement concrete structures These can divided into two main approaches: namely, the phenomenological and the micromechanics-based models The former had been inspired by the classical macroscopic theories of plasticity and damage, and attempts to account for gen- eral tri-axial states of stress and strain However, they have generally proven to be inadequate in providing unified constitutive relation that accurately reflect experimental data for arbitrary deformation histories.

To overcome these lacks, Bazant and Oh [9] introduced an alternative micro structural approach, which referred as ’microplane model’, based on slip theory of plasticity Over three decades, Bazantand his co-work have presented a se- ries of progressively improved version of the microplane models In particular,the M4 formulation of the microplane model has been proven to yield predic- tions that are in good agreement with experiment data It had been integrated into many commercial finite element (FE) code such as ATENA [18] and open source FE code as OOFEM Further more, many worldwide researchers imple- mented microplane M4 model in special FE softwares such as DYNA3D, ADINA, ABAQUS, LIMFES etc to solve their specific problems Many successfully ap- plications were reported in literature such asBhattachary[10],Baky[6],Liu and Foster[65, 66, 67],Heger et al [105].

In numerical study of this work, the microplane M4 constitutive material model was used for concrete material Moreover, to avoid mesh sensitivity, the crack band approach was also employed [8, 17] The microplane model M4 integrated in the ATENA software is according toBazant’s formulation [7], whose basic for- mula will be summarized in the next part Full details concerning the underlying hypothesis, basis relation of microplane M4 and advantage as well as disadvan- tage in practical applications can be found inBazant et al.[7, 15, 11],Babua et al.[5].

Through out numerical simulation framework, the finite element code-ATENA [18] was employed to carried out finite element analysis The program offers a wide range of options regarding element types, material behaviour and numer- ical solution controls etc The preparation of the input data (pre-processing) and evaluation of the numerical results (post-processing) are performed using the commercial program GID [24] These utilized advanced graphic user inter- faces features, auto-meshing as well as sophisticated post-processor and graphics presentation to speed up the analyses.

6.3.2 Microplane M4 material model in ATENA

With the constitutive law of the microplane model M4, the macro stress on the microplane is explicity determined from the stress-strain relationships, that have been developed for a generic microplane The micro stress are then combined using principle work to get macro stress at a point The micro stresses are split into normal and tangential on each microplane With the normal components further split into deviatoric and volumetric components Figure 6.2 shows the steps involved for extracting the macro-stresses at a point from the macro-strains.

The presented volume of material is viewed at the microstructural level, and is considered as three dimensional element defined by set of microplane of different arranged in regular patten Figure 6.2 depicts a typical representation, which includes a set of microplane 28 equally and distributed on surface of hemisphere.

These planes represent the damage or weak planes at the microstructural lever or plane of microcrack.

6.3 Microplane M4 material model for concrete 111

Figure 6.2.: Calculation macro stress scheme in microplane model

Figure 6.3.: Strain component on a micro plane

Finite Element Modelling

Previous studies on the bearing behaviour of steel-concrete composite with per- fobond dowels were mainly focused on experimental investigations with limited quantity.

In the Push-Out tests, both the yielding of embedded reinforcing bars in UHPC dowels and the local damage of concrete interspersing the holes have been found.

Moreover, the stress field in front cover is interesting to explain the contribution of transverse reinforcement in restricting the crack opening in the concrete sur- face As a simple mechanical model, the concrete interspersing the holes may be considered as a dowel loaded in shear and extreme local compression [62, 110].

However, the general validity of this model has not yet been sufficiently validated.

For the composite beams, the experimental study gave very meaningful informa- tion on ultimate bending capacity, failure mode and load slip distribution The strain development in specified section and local damage of concrete in the slab have been observed during test Among of beam tests, two of them were failed due to the collapse of the shear connectors, which occurred at very early as ex- pected Furthermore, the result indicate that, the load-slip behaviour of shear connector in composite beam are significantly different from the characteristic load-slip curve derived from Push-Out test Consequently, the distribution of the longitudinal shear force on each perforbon connector must be made clear.

Thus this is useful for enhancing the accuracy of the predicted bearing capacity of composite beams.

The chapter focus on the numerical modelling of Push-Out test and composite beams, by using the powerful ATENA software [18] (version 3.3.1) Fully three dimensional model with material nonlinearity was taken into account to evaluate not only global behaviour but also local deformation The chapter was divided into two parts: firstly, a finite element model for Push-Out test is developed,in order to predict the ultimate bearing capacity and explain the local damage of concrete zone as well as the yielding of steel The second part focuses on the development a 3D model for composite beams, with accounts for complex- ity geometry and nonlinear behaviour of materials The finite element analysis attempts to following aims: evaluate global structural behaviour of composite beams with slab made of UHPC, various profiles of steel girder and types of shear connector were considered with the same properties of UHPC slab. predict the ultimate load, performance of each steel profile and failure mode as well. determine the distribution of longitudinal shear force and the influence of the degree of shear connection on the bearing capacity.

In addition, this section also discussed several issues relating to computations such as, element mesh density, convergence, load control etc.

7.2 Modelling of Push Out Test

7.2.1 Finite element model Geometry of push-out specimens

Finite element model for SPOT was developed according to the experimental study, with most primary details and dimensions are the same testing specimens.

Some cosmetic details were ignored to reduce the potential risks in finite element mesh Consequently, this is also contributes to improved accuracy of the model.

Figure 7.1.: Geometry of push-out test specimens

7.2 Modelling of Push Out Test 129

The symmetry of the specimen was taken into account to reduce computational effort, therefore only a half of the specimens was modeled Figure 7.1 shows a half of the typical Push-Out test specimen, which was used in numerical investigation.

In the finite element model, the trapezoidal steel rip was replaced by rectangular plate with the same thickness.

The data input preparation and presentation analysis results were performed with ATENA/GID program A special tool, named Tool4Atena was developed to assist the preparation of the data file, to call ATENA solver module and to extract results.

Finite element mesh and kinematic conditions

Figure 7.2.: Finite Element model of Push-Out specimen

The Push-Out model was constructed with three solid blocks corresponding to steel, concrete slab and reinforcement parts The last two blocks were glued and meshed together, the nodes on the contact surfaces were merged completely The main aim is to ensuring that, the bond between UHPC and the reinforcement are perfectly without any relative slip Remaining steel part was meshed separately with concrete and reinforcing bars The interface between steel and concrete is processed in later The most important aspect for the Push-Out model is kinematic condition must be satisfied Which can be shortly described as follow: the perfobond rib can move in downward direction relatively with concrete slab under push-down load. the reaction force against moving down only appear at above half of concrete dowel. the interaction in normal direction of steel rib surface must be taken into account due to the deformation of the concrete block.

As shown in Fig 7.2, the FE model of Push-Out specimen was created with full three dimensional, which uses two types of elements from the ATENA element library The concrete slab, steel flange, perforated rib and reinforcing bar were modeled using CCIsoBrick8 3D three-dimensional solid elements.

There are two different idealizations which satisfy kinematic condition as stated above The consideration was based on economy of computation time, ease of data definition as well as the agreement with test data The first approach is to use constrain degree of freedom technique, which mergers the same displacement value in a degrees of freedom In fact, the coincident nodes at contact surfaces between steel and concrete are constrained with appropriated translation DOFs depending on their positions The second approach is to use gap element to represent the interface surfaces This method allows the transfer of force from concrete to steel and the friction force is also considered.

A comparison was carried out for several analytical cases For local damage investigation purposes, the coupling DOFs approach has more advantage It was adopted for Push-Out simulation, while interface approach was used for analysis of the composite beams.

Figure 7.3 shows the position of the DOFs to be constrained, where coincident nodes on a half above of dowel were merged completely with three DOFs Ux, Uy and Uz While at coincident concrete and steel rib surfaces nodes were coupled either in X or Z direction In addition those nodes at corner were couple in bothX and Z DOFs The merging of nodes replaces all the nodes that lie at the same coordinate location with only one node, and the lowest number of all the nodes merged is retained.

7.2 Modelling of Push Out Test 131

Because only of a half of the specimen was modeled, the appropriated boundary condition should be applied to the surfaces at the symmetric planes In fact, the restrain condition in X direction is assigned to all nodes on the symmetric plane.

And then, all nodes of bottom surface were restrained in Y direction Finally, the intersection lines of front and bottom surface were restrained for both Ux and Uy DOFs Fig 7.3 shows a typical boundary condition of a Push Out model.

Figure 7.3.: Loading, boundary conditions and constrain DOFs at contact surfaces between steel and concrete

As in actual tests, the prescribed displacement was applied at the top of the steel flange, all nodes have a uniform displacement in down ward direction The loading rate was increased very slowly by dividing into more than 250 sub steps.

In the elastic domain, the load steps was divided thinly scattered, otherwise in the inelastic domain (yielding or plastic/crushed) load step was specified veryfine Numerical experience indicates that, the control with difference load steps save more computational time and achieved better convergence during solving equilibrium equation system To improve convergence in each load step, linear search algorithm was turned on and a maximum of 50 iterations was allowed.

To obtain the load-slip response, the applied load can be calculated through by sum of vertical reaction forces at bottom of the concrete slab And the slip was measured as relative displacement between the nodes on the steel flange and that on the concrete slab, which are the same position in real test.

Conclusions and Future Perspective

UHPC is very promising new materials that are expected to find more appli- cations in the near future With UHPC the structures can be thiner, slender with daring new shapes and capable of carrying more heavy load as well as more durable in extreme conditions.

In this study the behaviour of composite beam made of UHPC with innovation continuous shear connector has been investigated The research described in this thesis comprise of two phases: phase I was an experimental investigation into the behaviour of Push-Out specimens and composite beams In phase II, numerical simulation was conducted to model the behaviour Furthermore a modelling approach based on Finite Element Code ATENA has been established to access the local behaviour.

The work undertaken in this thesis has identified a number of areas which needs further research It may be seen as “answers” to the questions raised in Section 1.3 Recommendations for further study are also given at the end of this chapter.

1 UHPC exhibits very high compressive strengths, which may reach 150MPa in normal curing condition and greater than 200 MPa if curing treatment is applied This material also shows high brittle behaviour, in stressed compression it could be explosive when crushed.

2 The tensile strength of UHPC is significantly higher than that of normal concrete The addition of steel fiber into concrete mixture improves re- markable the tensile strength Long fiber controls peak tensile stress while short fiber controls the post peak behaviour The cocktail fiber is more e ective Flexural strength can be reach 7.0 MPa to 25.0 MPa depend on volume of fiber content.

3 UHPC shows outstanding workability The slum flow varies from 65cm to 90cm The rheological properties of fresh UHPC are influenced by the concrete mixer design, mixing method as well as superplasticity The set time of UHPC is significantly delayed compared with normal concrete; final set does not occur until 12 to 24 hours after casting When setting is initiated, UHPC gains its compressive strength very fast.

4 Due to large amount of cement used, the shrinkage of UHPC must be taken into account for use.

5 The durability of UHPC are significantly better than that of normal concrete.

6 The fracture of UHPC is influenced by the fiber content and the casting direction as well as coarse aggregate The fracture energy varies in range of 5,000 to 20,000 N/m.

7 UHPC is a promising substitute for normal concrete and HPC in composite structures.

8 The high material cost is a restriction for the application of UHPC in practical engineering The finding of appropriated structural solutions for UHPC is still challenging for researchers and engineers as well.

8.1.2 Composite beam members made of UHPC under static load

1 UHPC enhances the performance of composite beam

The use of UHPC in composite beam lead to increased sti ness due to high elastic modulus The deflection of beam under service condition is much smaller comparison to conventional composite beams with NSC.

The ultimate strength is also higher allowing carry to more heavy load under critical conditions.

The high compressive strength of UHPC placed in compression zone optimize the load distribution in the section Thus increase signifi- cantly the load bearing capacity of composite member The size of member can be reduced.

Faster in strength development, higher in workability and not require- ment special curing condition lead to save more time, labor work, energy as well as the time to market.

2 The Steel-UHPC composite beams exhibit the same manner with conven- tional composite member with NSC

The structural behaviour of composite beam made of UHPC is similar to traditional composite beam in the three aspects: elastic, yielding and plastic domains In case of full shear connection, the simple rigid plastic can be used to predict ultimate plastic moment.

The continuous shear connection is able to e ecting transfer the load in composite beam with both Tee and I girders The longitudinal shear distribution is not uniform, the design for strength of shear connector must taken into account weakest connectors.

Composite section with Tee girder can provided 30% to 50% higher bearing capacity and sti ness than I section with the same cross sec- tion area of steel.

8.1.3 Perfobond based shear connectors in UHPC

1 The ductility of headed stud shear connector is significant in UHPC slab due to the deformation is restrained by very high strength concrete surrounding it The connector is often failed by shanked mode at the base This may reduce its fatigue strength under dynamic load The stud connector is not recommended to use in composite beam made of UHPC.

2 The perfobond shear connector exhibits good performance in shear load transfer as well as the ductility However, all the test data showed the characteristic slip uk is still smaller than 6.0mm as requireed for ductile connector Thus these connector can be only considered as non-ductile one in design.

3 The Perfobond connector without transverse reinforcement displays very poor ductility Thus it is not recommended in application The embedded rebar in dowel play a critical role for improving ductility of the connector.

4 Failure of connector is often caused by crushing of concrete and plastic of reinforcement rather than yielding of steel Therefore the ratio of cross section area of concrete dowel to steel rib should be adjusted to obtain appropriate load distribution between materials.

Appendices: Concrete mix proportional

A.1 List of tables for constituent materials

UHPC-B4Q 1% steel fiberUHPC-G7 1% steel fiberUHPC-G7 0.5% steel fiber

A.1 List of tables for constituent materials 167

Appendices: Standard Push-Out Test

B.1 Experimental results of Standard Push-Out test

Series 1: Headed stud ( 16mm, Bst500), test setup S1 Series 2: ODW without rebar, test setup S2

Series 3: ODW with rebar in core, test setup S2 Series 4: ODW with rebar in core and cover, test setup S2 Series 5: CDW without rebar, test setup S1

Series 6: CDW with rebar in core, test setup S1 Series 7: CDW with rebar in core and cover, test setup S1 Series 8: CDW with rebar in cover, 0.5% steel fiber, test setup S1 Series 9: CDW with rebar in cover, 1.0% steel fiber, test setup S1

Series 10: CDW with 8mm rebar in core and cover, 0.5% steel fiber, test setup S1

Series 11: CDW with 12mm rebar in core and cover, 0.5% steel fiber, test setup S1

B.2 List of drawings and charts

Push-Out test setupPush-Out rebars arangementChart of Load-slip and crack openning

Figure B.1.: Push-Out test setup S1 and S2

B.2 List of drawings and charts 171

Figure B.2.: Rebars arrangement of Push-Out specimens

(c) Figure B.3.: Push-Out test reults: Load-Slip and Crack opening, Series 1-Headed stud shear connector, specimen-1(a), specimen-2(b), specimen-3(c)

B.2 List of drawings and charts 173

Figure B.4.: Push-Out test reults: Load-Slip, Series 2-ODW without rebar (left), Series 3-ODW with rebar in core(right)