This may be the author’s version of a work that was submitted/acceptedfor publication in the following source:

Elchalakani, Mohamed, Hassanein, Mostafa Fahmi, Karrech, Ali, Fawzia,Sabrina, Yang, Bo, & Patel, Vipul(2018)Experimental tests and design of rubberised concrete-filled double skincircular tubular short columns.Structures, 15, pp. 196-210.

This file was downloaded from: https://eprints.qut.edu.au/123300/

c© Consult author(s) regarding copyright matters

This work is covered by copyright. Unless the document is being made available under aCreative Commons Licence, you must assume that re-use is limited to personal use andthat permission from the copyright owner must be obtained for all other uses. If the docu-ment is available under a Creative Commons License (or other specified license) then referto the Licence for details of permitted re-use. It is a condition of access that users recog-nise and abide by the legal requirements associated with these rights. If you believe thatthis work infringes copyright please provide details by email to qut.copyright@qut.edu.au

License: Creative Commons: Attribution-Noncommercial-No DerivativeWorks 2.5

Notice: Please note that this document may not be the Version of Record(i.e. published version) of the work. Author manuscript versions (as Sub-mitted for peer review or as Accepted for publication after peer review) canbe identified by an absence of publisher branding and/or typeset appear-ance. If there is any doubt, please refer to the published source.

https://doi.org/10.1016/j.istruc.2018.07.004

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

EXPERIMENTAL TESTS AND DESIGN OF RUBBERISED CONCRETE-FILLED DOUBLE SKIN CIRCULAR TUBULAR SHORT COLUMNS

MOHAMED ELCHALAKANI *1, M. F. HASSANEIN 2, ALI KARRECH1, SABRINA FAWZIA3,

BO YANG4, V. I. Patel5

1Department of Civil Engineering, Faculty of Engineering, Computing and Mathematics, The University of Western Australia, Australia

2 Department of Structural Engineering, Faculty of Engineering, Tanta University, Tanta, Egypt

3School of Civil Engineering and Built Environment, Science and Engineering Faculty, Queensland University of Technology (QUT), 2 George Street, Brisbane, QLD 4000,

Australia

4School of Civil Engineering, Chongqing University, Chongqing 400045, China

5 School of Engineering and Mathematical Sciences, La Trobe University, Bendigo, VIC 3552, Australia

ABSTRACT

The adequacy of rubberised concrete (RuC) for use in structural columns is, currently,

investigated experimentally through the use of cold-formed double-skin circular steel tube

confinement. The RuC is of particular interest because the aggregate can be sourced from

recycled tyres, so it is a form of sustainable concrete, and it possesses superior mechanical

properties to conventional concrete such as increased ductility and energy absorption.

Rubberised concrete does have one major issue in that it has a low compressive strength

compared to normal concrete, which limits its application. The aim of this study is to evaluate

the effectiveness of confinement in overcoming this mechanical deficiency by using

rubberised concrete-filled double-skin tubes (RuCFDST). The experimental program

involves testing and measurements of key mechanical properties including compressive

strength, hoop and axial strains, and compressive load-deflection curves. A total of 15

composite specimens were examined to ascertain the varying properties of single-skin,

double-skin, confined, unconfined, standard and rubberised concrete. Recycled rubber * Corresponding author (Mobile: +61479199629; Fax: +614864882376; E-mail:

mohamed.elchalakani@uwa.edu.au).

*ManuscriptClick here to view linked References

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

2

particles ranging from two to seven millimetres in size were used to replace 15% and 30% of

the fine and coarse concrete aggregate by weight. The rubber particles were treated with

sodium hydroxide solution resulting in increased bonding strength to the concrete. Circular

hollow section (CHS) configurations of different internal and external dimensions were also

examined to further understand the mechanics of double-skin confinement. Confinement in

RuCFDST showed significant improvements in strength and ductility properties.

Experimental results proved to be in agreement with design ultimate axial strength

predictions proposed by existing methods and design codes. Given the exceptional ductility,

energy dissipation and improved strength of RuCFDST, this study shows the potential

viability of RuCFDST as structural columns particularly in areas that are prone to seismic

activity.

KEYWORDS

Rubberised concrete; Concrete-filled double skin steel tubes; Stub column; Confinement;

Axial compression; Design model.

1. INTRODUCTION

The need for environmentally sustainable and economical construction materials has been

growing rapidly in recent years. There has been a heavy focus on creating new composite

materials that can utilise the advantages of different mechanical properties, optimise

construction costs and minimise environmental pollution [1]. As a direct result, rubberised

concrete (RuC) has been rapidly gaining popularity. Disposal of scrap tyres are a major

environmental issue around the world with approximately 50 million end-of-life tyres being

discarded each year in Australia alone [2]. Disposal of rubber tyres into landfills can cause

major environmental issues such as creating an ideal breeding environment for rats and pests

in addition to posing potential fire hazards [3]. This has consequently sparked interest in the

use of rubber particles from waste tyres as replacement aggregate in concrete. This solution

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

3

not only promotes the recycling of scrap tyres but also reduces the need for further natural

aggregate extraction. Extensive research shows improved ductility, dynamic energy

absorption and post-failure compressive loading for rubberised concrete. Unfortunately, these

unique characteristics come with a major drawback, which is the significant reduction in the

concrete's compressive strength [4]. Loss in strength can be as much as 80% compared to

standard concrete depending on the percentage replacement and size of the rubber aggregate

[4-5]. Accordingly, efforts have been made to improve the strength of rubberised concrete by

enhancing the mixture by using additives and pre-treatment of the rubber particles to improve

rubber-concrete bonding [1-6]. Although results have shown improvements in compressive

strength, the percentage increase in strength to date has not been sufficient to justify the use

of RuC in structural members.

The mechanical properties and behaviour of RuC, in which a portion of the concrete

aggregate is replaced by rubber particles, have been well documented. Topçu, Toutanji and

Fattuhi [5, 7-8] deduced that the addition of rubber aggregate (RA) changed the behaviour of

plain concrete from a brittle failure, to a relatively ductile, plastic failure, which could absorb

large amounts of energy under compressive and tensile loading. The RuC specimen could

also withstand significant post-failure loading and displacement without full disintegration.

Segre and Joekes [9] conducted toughness testing of flexural specimens and discovered that

high toughness was shown in specimens that contained RA. Khatib and Bayomy’s [10]

analysis suggested that the reduction in slump values at larger volumes of rubber content

increased the air content and decreased the unit weight of the specimen.

On the other hand, concrete-filled steel tubes (CFST) have been a very popular topic of

research around the world. CFSTs were first introduced by Naka and Kato [6] to improve

compressive members used in power transmission towers. Morino and Tsuda’s [11]

investigation found that CFSTs provided improved compressive strength and ductility

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

4

properties compared to steel and reinforced concrete systems. This was due to the restraint

provided by concrete to delay local buckling of steel, and the increase of strength due to

confinement. Other materials used for concrete confinement have produced similar results to

that of steel such as crumb rubber concrete confined by fibre reinforced polymer (FRP) tubes

[12]. Youssf et al. [12] found that using 3 layers of FRP confinement largely improved the

compressive strength of crumbed rubber concrete by 186%. The most notable literature on

RuC confinement by steel tubes is found in Duarte et al. [13]. Duarte et al. [13] investigated

the effect on compressive strength of rubber concrete using circular hollow section (CHS)

and square hollow section (SHS) tubes. It was seen that confinement did indeed increase

compressive strength, with CHS confinement found to have more significant improvement

compared to SHS confinement. Since the introduction of CFSTs, further research has been

made on alternative concrete confinement systems which have shaped the idea of concrete-

filled double-skin tubes (CFDST) [8]. CFDST technology [14-16] is the basis of this

investigation, with an aim to improve rubberised concrete compressive strength by

introducing double-skin confinement. The confinement consists of two circular hollow

sections (CHS), one on the inside and one on the outside, with the annulus being filled with

rubberised concrete as shown in Fig. 1. This configuration is of particular interest, as

previous studies have concluded that CHS performance in improving ductility properties and

confinement are much improved compared to its square hollow section (SHS) counterpart

[13]. An experimental investigation will be conducted to establish a comprehensive

understanding of rubberised concrete filled double skin tubes (RuCFDST). The experimental

analysis will particularly focus on measuring compressive strength and deformation under

static loading. The loading on the specimens will ideally model the compressive loading on

structural members such as bridge piers.

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

5

The fundamental goal of this investigation is to discover new techniques to improve

rubberised concrete axial compressive strength, to develop its feasibility for use in structural

members, predominantly in seismic prone regions. This investigation will consider the

viability of using double-skin steel confinement, with a circular hollow section (CHS) inner

and outer configuration to improve RuC mechanical properties. Three CFST and twelve

CFDST confined rubberised and normal concrete specimens will be tested, with crumbed

rubber replacing fine and course aggregate at 0, 15 and 30% of total aggregate by weight.

Effects on ductility and compressive strength will be assessed for each test specimen.

Previous postulated ultimate compressive axial strength predictions by Zhao et al. [17], Tao

et al. [18], Hassanein and Kharoob [14] and Euro Code 4 (EC4) [19] will be evaluated and

verified by the experiment findings. It is hoped that this investigation provides a building

block for further study in rubberised concrete strength improvement.

2. MATERIALS AND METHODS

2.1 Mixture material properties

2.1.1 Steel

Cold-formed steel manufactured in accordance to AS1163 [20] was used in the construction

of the specimens, which were delivered by Metalcorp Steel (Perth, WA). The 165.1 mm outer

diameter circular tubes were coated on both sides with electro-galvanised zinc coating. All

other circular hollow section steel tubes were plain black painted and were not galvanised.

The CHS is grade CL350L0 with a nominal yield stress of 350 MPa. A summary of the

dimensions and material properties of the steel sections used to produce the specimens can be

found in Table 1. , in Table 1, is defined as the circular hollow section slenderness taken

from Clause 5.2.2 AS4100 [21].

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

6

(1)

2.1.2 Cement

General purpose Portland cement compliant with AS3972 [22] was acquired from Swan

Cement and used as the binder material in the normal and rubberised concrete mixes. The

chemical composition of the cement is shown in Table 2.

2.1.3 Choice of normal/rubber aggregate sizes

Since the smallest specimen annulus size was 68 mm, a maximum aggregate size of 7 mm

was chosen to achieve sufficient compaction and homogenous particle distribution. The

crumbed rubber particles were delivered by Tyrecyle (New South Wales), which is

Australia’s leading national tyre recycler. The delivery consisted of two sets of rubber sizes,

2-5 mm, and 5-10 mm. The 5-10 mm particles were put through a 6.75 mm sieve to attain a

maximum aggregate size of 7 mm. A visual representation of the crumb rubber particles sizes

is shown in Fig. 2.

2.1.4 Particle size distribution of aggregates

Sieve testing is essential for understanding the particle distribution used for aggregate

replacement. This test enables categorisation of rubber particles into poor/well-graded types.

AS3638 [23] sieving procedures were followed to obtain a particle size distribution (PSD) of

the concrete mixture material. Results from sieve testing can be found in the PSD provided in

Fig. 3. All particles except for the 5-7 mm aggregate were observed to be well graded. The

poor grading of these particles was possibly due to the nature of sieving the 5-10 mm

aggregates to a maximum size of 7 mm.

2.2 Rubber pre-treatment

Given the lack of adhesion at the rubber-cement interface of RuC [4], Segre and Joekes [9]

conducted treatment on the rubber particles with a saturated NaOH aqueous solution. Using

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

7

electron microscopic examination, treated rubber was observed to have enhanced rubber-

matrix adhesion, which was further confirmed by the reduction of water absorption,

improvement in mechanical properties and abrasion resistance. Li et al. [24] conducted

testing using surface treatment and physical anchorage of shredded rubber. Their finding

presented that rubber fibre had an improved performance over chipped rubber and that

residual steel belt wires from truck tyres had positive effects on strength. Güneyisi et al. [25]

found improved compressive strength could be achieved by incorporating silica fumes into

the RuC matrix. Ho et al. [26] studies discovered that the use of RuC improved fracture

properties. Recent studies analysed the impact of replacing sand with RA in self-compacting

concrete (SCC) [1, 27-29]. SCC’s pronounced powder content and dense, high-compact

nature results in increased strength and brittleness. SCC can be also used in RuC applications

that require higher workability.

Previous experimental investigations [23-24] found that pre-treatment has a significant

impact in improving the adhesion at the rubber-cement surface interface. It is proposed that

the low adhesion levels are due to several factors such as: (1) waste rubber zinc stearate

coating creating a soapy layer that repels water; (2) low hydraulic conductivity; and (3) the

smooth surface interface of rubber [12]. The most commonly used chemical for rubber pre-

treatment is sodium hydroxide (NaOH) solution. Therefore, this is what has been used in the

present study. NaOH pre-treatment has proven to improve compressive strength by up to 25%

[30]. This is achieved by removing the unwanted zinc stearate layer and also creating a

rougher surface on the rubber particles [23, 31].

Prior to treatment, the rubber particles need to be washed to remove impurities, organic

matter and dust that can weaken the bonding strength at the rubber-cement interface [23]. The

rubber is then soaked in a 10% NaOH aqueous solution for an optimal duration of 24 hours

suggested by Mohammadi [32]. The solution is then drained and the rubber is rinsed with

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

8

water until a pH of 7 is achieved. Finally, the rubber is semi saturated through a water

soaking process to increase the specific gravity of the rubber, thereby reducing the floating

tendency of the lightweight rubber particles. This process ultimately results in a stronger and

more homogenous rubber-concrete mixture.

2.3 Mix procedure/design

The concrete mix procedure is critical to achieving good quality concrete. Too much mixing

can cause the mixture to separate due to differences in size and density (concrete

segregation), whereas too little mixing does not allow sufficient hydration and mixing of the

different particle sizes and densities in the concrete. Introducing rubber particles to the mix

amplifies segregation due to its low density compared to other concrete constituents. To

reduce segregation, pre-treatment as explained in Section 2.2 is used as well as timed mixing

techniques which have been adapted from previous literature. Rubber’s relatively low

specific gravity causes other issues such as the migration of light rubber particles to the top of

the concrete due to the vibration processes. To minimise rubber flotation and air voids,

manual rodding compaction is used as an alternative compaction process to minimise rubber

flotation while achieving adequate compaction (removing air voids in the concrete mix).

Slump tests were performed prior to pouring in an attempt to further ensure that the required

workability and consistency needed to compact the concrete mixture was met. It was found

that a slump of 150-170 mm was optimal for the specimens used, with slumps above 170 mm

showing formation of undesired segregation, and slumps values under 150 mm being too

difficult to compact. Small amounts of superplasticiser solution were added to the mix when

slumps of at least 150 mm were not achieved to allow for greater workability.

Three concrete mix batches were produced: 0% (control), 15% and 30% rubber replacement

to aggregate by weight. After careful consideration of previous literature on mixing

procedures and different methods of mixing, the following steps were used to produce the

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

9

optimal concrete: (1) mix all dry aggregate for 1 minute; (2) add 10% of the total water and

mix for 1 minute (for RuC add rubber particles at this stage); (3) add cement and mix for 1

minute; (4) add half of the remaining water gradually and mix for 1 minute; (5) add the

remaining half and mix for another 1 minute; and (6) finally small amounts of super-

plasticiser was gradually added and mixed for 1 minute until a slump of 150-170 mm is

achieved.

It is important to note that the rate at which water was added to the mix had a large impact on

the slump and mixing of the concrete. Gradually adding water to the mixer produced the best

results as it reduces clumping of material and made sure that the mixture was consistent and

homogenous throughout. In addition, a water-to-cement ratio (w/c) of 0.5 was used to achieve

the desired slump values. The water absorbed by the rubber through the water soaking

process was accounted for when producing RuC by reducing the overall water required for

the mix by the amount absorbed by the rubber itself. This was done to ensure the 0%, 15%

and 30% rubber replacement were achieved consistently. Pouring times between different

mixtures were set for timeframes of similar external temperature and humidity conditions.

Fig. 4 shows the distribution of aggregates within the mixture. The rubber aggregates can be

seen to successfully distribute homogenously through the cross-sectional area of the trial

samples.

2.4 Concrete strength

Three test trials for each concrete mix, 0%, 15% and 30% rubber replacement by weight were

performed to calculate the mean concrete strength. Concrete cylinders 100 mm in diameter

and 200 mm long were prepared and tested in a 600 kN capacity Baldwin Machine to

AS1012.9 [33] on the day when 28-day curing was achieved. The density of each sample was

also measured prior to compressive testing. It was seen that the 15% and 30% rubber

replacement by total aggregate weight decreased the density by approximately 8% and 14.5%

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

10

respectively. A summary of the concrete compressive tests and density measurements are

shown in Table 3. Note the convention used in the table: CT-XX-XX, Compression test - %

rubber replacement - trial number. Sample CT-15-02 is disregarded in averaging due to visual

imperfection in compaction of the concrete, which significantly lowered the concrete

strength.

3 TEST PROGRAM

3.1 Specimens

Two sets of testing were conducted to find the compressive strength of empty steel tubes and

confined concrete (CFST and CFDST). These tests were conducted to define the behaviour of

confinement in improving the overall concrete strength compared to its standalone

counterpart. A total of ten specimens were used for empty steel tube testing and fifteen

specimens for confined concrete testing. All samples were cut to length (L=400 mm) for each

corresponding test specimen. The CFST and CFDST specimens were capped at one end by

tack welding a 10 mm thick square steel plate to allow for casting of concrete. A summary of

CFST and CFDST empty specimen properties can be found in Table 4. There is a total of 3 of

each specimen configuration from Table 4, filled with concrete mixture of 0%, 15% and 30%

rubber replacement to total aggregate.

3.2 Concrete preparation

All concrete samples were compacted using manual rodding. Samples that contained concrete

were placed into a humid room immediately after pouring was completed to minimise

concrete shrinkage. The samples remained in the humid room for 21 days, then removed and

placed in an undercover room for 7 more days to complete the 28-day concrete curing

process. The confined specimens experienced some minor amounts of shrinkage. Since the

top surface was no longer perfectly flat due to shrinkage, a thin layer of grout was applied to

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

11

each specimen 2 days prior to completion of curing to ensure a flat surface for even loading

throughout the steel-concrete composite.

3.3 Test procedure

3.3.1 Empty circular steel tubes testing

Two trials of each tube diameter shown in Table 1, with 150 mm nominal lengths were tested

using a 600 kN Baldwin machine with a rate of displacement less than 1 mm/minute. Both

load and displacement readings were read from the machine. Real-time data was recorded on

a computer using a data logger. The galvanised cylinders with 165.1 mm outer diameter were

tested using a 5000 kN capacity DLS500 machine, since the capacity of the cylinders

exceeded 600 kN. A linear variable differential transformer (LVDT) was used in this case to

measure the axial displacement of the specimen.

3.3.2 CFST and CFDST testing

A load control of 1 kN per second was used to test the confined composite columns with a

rate of displacement less than 0.5 mm/minute. One trial of each specimen configuration and

rubber replacement volume was tested for axial compressive strength using a 5000 kN

capacity DLS 500 machine. The loading from the machinery was captured using a 3000 kN

NATA approved load cell placed on top of the specimen. Two linear wire position

transducers (string pots) were attached to the top and bottom platens at each corner to record

axial shortening. Two strain gauges were glued to the specimen vertically and horizontally to

measure hoop and axial strain. A data logger was used to transfer the load, displacement and

strain gauge measurements to a computer for the duration of the testing. The specimen and

the load cell were placed concentric with the bottom and top fixed platen to ensure centric

loading. Fig. 5 illustrates the test setup.

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

12

4 TEST RESULTS OF EMPTY CHSS

Short circular hollow section (CHS) steel tubes subjected to uniform axial compression

buckle either elastically or plastically depending on the diameter to thickness ratio, D/t. In

general, thin-walled cylinders buckle in a so-called diamond mode, whereas thick walled

cylinders buckle in the axisymmetric mode (also known as elephant mode) at one end of the

tube [21]. The D/t ratio distinguishing the two modes of buckling failure can be largely

dependent on residual stresses and imperfections in the cylinders. An estimated D/t ratio of

less than 40 is found to produce elephant mode buckling failure [17]. Thus, this buckling

mechanism is the expected mode of failure for the tubes in this study as they satisfy the D/t <

50 criterion. Ultimately, the results from these tests were used to quantify the material

properties and how the different steel sections affect the overall strength of CFST and

CFDST tubes.

The load deflection curves from axially loaded CHS tubes for each trial are shown in Fig. 6.

It should be noted that the unloading process in C1, C2, C3 and C4 type specimens were seen

to be rather violent from to the sudden release of the axial loading. This behaviour could not

be controlled because of the nature of the machinery not allowing for controlled unloading.

On the contrary the C5 type specimens showed gradual unloading which is more

representative of the unloading behaviour of the material using the DLS 500 machine. The

results from both trials showed almost identical load-deflection curves and ultimate peak

loads (Pu). Once buckling failure was initiated, the axial shortening rapidly increased and

then decelerated as the upper and lower fold faces came closer to touching. The smaller

diameter cylindrical tubes exhibited a flatter load-deflection curve with a later peak loading

point, which may be due to a smaller and more gradual buckling caused by its smaller section

slenderness λso. The larger λso sections displayed a steeper reduction in post-failure loading,

like that of Zhao’s findings [18]. From the equal outer diameter specimens, C3 and C4, we

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

13

can deduce that an increased thickness results in a higher ultimate peak load with an almost

identical post-failure mechanism. This result is expected as a larger gross area of steel can

withhold a larger applied axial load. The galvanised coated steel, C5, showed a larger

Young’s modulus compared to all other sections. There is also a distinct kink in the

galvanised steel load-deflection curve, which could be a result of the galvanisation of steel

causing the cylinder to become more brittle due to strain-aging. A summary of the ultimate

peak loads for each trial and an average peak load for each specimen type can be found in

Table 5. Furthermore, an axisymmetric mode of failure was confirmed to be the buckling

mode of failure which occurred at the top or bottom of each specimen. The typical failure

modes obtained for CHS specimens in these tests are showed in Figs. 7(a),(b) for two series

of tests.

5 TEST RESULTS OF CFSTS AND CFDSTS

There is a total of three CFST and twelve CFDST specimens tested. The single skin testing

was conducted as a base case used to compare the differences between double-skin and

single-skin confinement. The single-skin confinement testing was conducted only on C5 steel

sections with 0%, 15% and 30% rubber replacement.

5.1 Test results of CFST and RuCFST columns

The load-deflection curve of the single-skin confined concrete is shown in Fig. 8. The naming

convention is as follows: circular hollow section (CHS) – outside steel (O) diameter –

percentage of rubber particle replacement. Only the C5 outer diameter (165.1 mm outer

diameter and 3.5 mm thick) galvanised steel tube was tested for single-skin concrete

confinement. It should be noted that the sudden load drop in the 15% rubberised concrete’s

load-deflection curve is due to a load cell disconnection and does not represent the unloading

of the specimen. A summary of peak loads can be found in Table 6.

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

14

The results show that the peak ultimate axial load is significantly decreased with an increase

in rubber particle replacement confirming the same results found by Refs. [13, 34-35]. This is

to be expected as concrete compressive strength significantly decreases with increased

replacement of rubber particles (refer to Section 2.4). The standard 0% concrete presents a

more pronounced peak load with a significant drop in compressive loading post initial

buckling failure compared to the 15% and 30% rubber replacement. This phenomenon may

be explained by analysing the Poisson’s ratio of rubberised concrete compared to standard

concrete. Rubberised concrete has a Poisson’s ratio closer to that of steel than standard

concrete [31]. Thus, the transverse elongation matches the steel more closely and potentially

results in the buckle fold filling with concrete at a faster rate than that of standard concrete,

which has a lower Poisson’s ratio. The filling of concrete within the buckle fold continues

until the buckled section where the overall specimen becomes stiff enough to withstand

further loading. Standard concrete therefore has a larger dip in carrying load post failure, as it

takes a longer duration to fill the void created by the buckling of the steel.

5.2 Test results of CFDST & RuCFDST columns

A total of four different double-skin configurations were tested for axial strength at 0%, 15%

and 30% rubber replacement. Load-deflection curves for each configuration are found in

Fig. 9, 10, 11 and 12. The naming convention in the load-deflection curves are as follows:

circular hollow section (CHS) – outside steel (O) diameter – (outer cylinder thickness if

applicable) – inside steel (I) diameter – percentage of rubber particle replacement. Table 6

summarises the peak axial force post initial buckling. These values are used for predictive

design capacity comparisons following Zhao’s [17], Tao’s [18], Hassanein's [14] and

Eurocode [19] design models. The numbers in brackets are the percentage reductions in

strengths due to the rubber particles.

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

15

As expected an increase in rubber content in the concrete resulted in a lower ultimate peak

load for all test cases. CHS-O165-I42 and CHS-O165-I89 are seen to behave like its single-

skin counterpart CHS-O165 found in Section 4. This is due to the small internal steel core

diameter skewing the specimen to behave more as a single-skin specimen rather than a

double-skin specimen. Normal concrete holds approximately the same axial loading in both

single-skin (CHS-O165) and double-skin configuration (CHS-O165-I42 and CHS-O165-I89).

This confirms the previous results of Zhao and Han [36] who found that CFDST columns

nearly have similar performance to traditional CFST columns of the same dimensions of

outer steel tube and strength of materials, but with less weight due to the void inside the inner

tube. CHS-O165-I42 specimens experienced minor flexural buckling once the first local

buckles were formed due to the slenderness of specimens. The flexural buckling was found to

have negligible effects on peak axial load on the specimen. In general, the peak load for

rubberised concrete occurred at later axial shortening displacements. This is believed to be

due to the RuC forming micro-cracks and pushing against the steel boundary in a ductile

manner causing a plateau in loading. However, normal concrete failure is caused by the

formation of large crack propagating down the concrete and it is expected that the post peak

dip in load would be greater than RuC.

It can interestingly be observed, from Table 6, that the single-skin specimen CHS-O165 with

15% rubber content and the double-skin specimen CHS-O165-I42 with 30% rubber content

have nearly the same axial load. Hence, it can be concluded that using double-skin specimens

cannot only decrease the weight of the column [36], but it also can be more environmentally

friendly through consuming much more rubber content without a reduction in the strength

compared with the single-skin specimens of the same diameter.

5.3 Deformed shapes of normal and rubberised CFSTs and CFDSTs

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

16

A family photo of pre- and post-deformed shapes for column CHS-O165-I89-30 is presented

in Fig. 13. The outwards buckling generally occurred in the mid-height of the column as

obvious from the figure. The internal buckling failure of the CHS-O165-89-30 specimen is

illustrated in Fig. 14. Fig. 15 shows a lapsed photo of the CHS-O165-89-30 specimen sample

corresponding to the load-deflection curve where outwards buckling becomes evident.

Generally, buckles become fully formed at a point approximately half way between

subsequent crests and troughs of the load-deflection curve. Appendix 1 displays a progressive

compression of the specimens with increasing axial deflection, further illustrating the

buckling failure.

5.4 Concrete-steel bonding

The specimens have a linear elastic behaviour of about 60-70% of the initial peak load when

a full bond between the concrete and the steel surface does exist. Once this limit is exceeded,

non-linear behaviour becomes prominent as the steel tubes begin to yield and concrete

subsequently presses against the inner and outer tubes non-linearly. Micro-cracks start to

emerge in the concrete during this stage and the buckling of the inner and outer steel tubes

produce fill voids. Instead of a brittle concrete failure, the concrete behaves in a more ductile

manner as the geometry of the steel confines it.

As further axial shortening of the specimen takes place, more concrete fills the initial buckle

causing it to regain some of bonding at the steel-concrete surface due to the high pressures at

the interface. This fundamental behaviour is the one of the core advantages of using CFSTs

and CFDSTs, as they can withhold incredible loads post failure.

Segment cut-outs of the 0%, 15% and 30% concrete at buckled locations are shown in Fig. 16

It was seen that normal concrete had a rigid bond at the buckled area but the 15% and 30%

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

17

rubberised concrete had a comparatively weak bonding performance. The rubberised concrete

mixture had segregated within the composite, becoming crumbly and easy to move around.

This failure is related to the high internal stresses induced perpendicular to the axial load

compression by the low modulus of elasticity of rubber particles. This causes cracks to form

around rubber particles creating a separation of the materials. The separation is further

amplified from the use of high water to cement ratio of 0.5 in the mixing procedure. Fig. 17

better depicts the crumbly nature of rubberised concrete post-failure.

5.5 Exterior Steel Strain Gauge Data

Hoop strain and axial strain measurements for all confined specimens were measured using 5

mm long strain gauges. Fig. 18 shows the relationship between axial compressive load and

measured axial strain (positive compressive strain), in addition to hoop strain (negative

tensile strain) for CHS-O165-I89 configurations. The load verses strain plot in Fig. 18 can be

divided into three main components. The first is the elastic component, which continues until

concrete unconfined strength is reached. Since normal concrete and steel generally have a

Poisson’s ratio of around 0.15 and 0.3 respectively, the concrete requires cracking first before

it can feel the effects of steel confinement. This occurs in the nonlinear component of the

curve as the concrete cracks and presses against the inner and outer walls. The final linear

component is where the steel and concrete start expanding at roughly constant rate once

initial cracks have developed.

5.6 Ductility and Energy Absorption

A ductility index (DI) analysis of each specimen was conducted to quantify the ductility of

each specimen. The DI is the ratio A1/A2, where A2 is the recoverable elastic energy at

initial peak load failure and A1 is the irrecoverable plastic energy as seen in Fig. 19 [13].

Specimens with larger DI values are more ductile. The DI index of all specimens is shown

graphically in Fig. 20. Specimens with 0% rubber replacement are shown to be significantly

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

18

less ductile than their RuC counterparts. The results show that RuC has up to 2.5 times better

ductility with respect to normal concrete CFST and CFDST. The ductility performance of

15% and 30% rubber replacement are similar in most cases. It is also seen that double-skin

confinement with smaller ratios are generally more ductile. This may be due to the

concrete being able to push a large surface area of both inner and outer steel sections

allowing for a more ductile progressive failure.

6 PREDICTION OF CFDST AND CFST COLUMN STRENGTH

6.1 Design model by Zhao and Grzebieta [17]

The CFDST and CFST ultimate strength can be estimated as the superposition of the section

capacities of the concrete core, outer steel tube and inner steel tube. This is formulated as:

(2)

Where can be found in Section 4.1.2 and is given by:

(3)

And, (4)

Where, is the outer diameter of the concrete core, and is the outer diameter of the inner

steel. is equal to 0 for CFSTs.

6.2 Design model by Tao and Han [18]

Tao and Han [18] adopted a similar prediction approach in which the capacity section was

superimposed to find the composite sections ultimate strength. The proposed calculation is as

follows:

(5)

Where is the capacity of the outer steel with the sandwich concrete and is the

compressive capacity of the inner steel tube computed as ; and are the inner

steel cross-sectional area and yield strength respectively. is given as:

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

19

(6)

In which,

(7)

(8)

Where:

is the cross-sectional area of the outer steel

is the cross-sectional area of the sandwiched concrete

is the yield strength of the outer steel tube

is the characteristic concrete strength given by: 0.67 , where is the

characteristic cube strength of concrete

is the confinement factor given by:

is the nominal cross section area of concrete, given by: where,

and are the outer steel diameter and wall thickness respectively

is the hollow section ratio, given by:

is a coefficient given as: , where is the steel ratio given by

is a coefficient given as: , where is the

nominal steel ratio given by

6.3 Design model by Hassanein and Kharoob [14]

A unified axial load bearing capacity for circular CFDST columns was recently proposed by

Hassanein and Kharoob [14] for 150/ etD based on the models suggested by Yu et al [37]

for solid and hollow circular CFST columns, as follows:

sisyiscccsesyeHasul AfAfAfP1

3.01, (9)

Where:

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

20

is the confinement coefficient, cksy ff /

is the steel ratio, cs AA /

is the solid ratio, )/( kcc AAA

sA is the area of the external steel tube,

cA is the area of the concrete core,

kA is the area of the hollow part,

6.4 Design model by Eurocode 4 (EC4) [19]

Unfortunately, at this present time EC4 does not provide design considerations for CFDST

columns. However, EC4 provides compressive design formula for estimating CFST ultimate

compressive strength. CFST design strength considering confinement effects is given in

clause 6.7.3.2(6) of EC4 (given no additional steel reinforcement bars and assuming no

eccentricity) as:

(10)

Where:

is the reduction factor for the outer steel tube given by:

(but

is the concrete enhancement factor due to confinement effects given by:

(but and;

is the relative slenderness given by: (but

is the critical buckling load, used in calculation of slenderness parameter given by:

Is the effective flexural stiffness of the cross-section given as:

with a correction factor

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

21

Pagoulatou et al. [38] proposed that the CFST EC4 equation for CHS can be modified to

yield a design formula for CFDSTs. To achieve the best approximation, consideration of

confinement effects was only taken at the outer steel tube. This is to close agreement to the

experimental results as inner CHS did not provide significant confinement to the concrete

core. Thus, the modified equation for CFDST is as follows:

(11)

6.5 Calculated strengths and discussion

Table 7 shows the comparison of experimental to theoretical ultimate peak strength for

CFDST and RuCFDST using the models in the preceding sections. Zhao and Grzebieta [17]

strength model provides the most conservative design values, underestimating by up to 17%.

This result is to be expected as the model does not account for additional strength provided

by confinement effects to the concrete by the steel sections. An interesting result is that

rubberised concrete seems to have better confinement in double-skin configuration compared

to its normal concrete counterpart using Zhao’s strength predictions [39-42]. Tao and Han

[18] predictions are mostly good with the average of the predictions being close to unity.

However, this model can slightly overestimate the strength as seen in specimen CHS-O114-

3.2-I42-15 by 5% in maximum. Hassanein and Kharoob [14] strength predictions, on the

other hand, show better average strength ratios though this suggestion provides unsafe results

up to 11% in case of 30% rubber content. Although on average, the modified EC4 model

estimate is near unity (99%) it appears to overestimate CFDST with rubberised concrete with

a large variability in values across all specimens. Given that Zhao’s model provides highly

conservative design capacities of the column, it is recommended to choose the most suitable

results from Tao’s formula for design.

7 CONCLUSIONS

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

22

This paper presents the results of an experimental investigation on the strength and ductility

of CFST and CFDST filled with normal and rubberised concrete (RuC). The main

conclusions of this study are summarised in the following points:

1. RuC mixes have a lower compressive strength than normal concrete. Concrete strength

decreased by 50% and 79% for 15% and 30% rubber replacement by aggregate

respectively. The 15% rubber replacement with 25Mpa strength is a viable alternative for

applications such as footpaths and footings.

2. NaOH rubber pre-treatment improved the concrete rubber bonding and reduced concrete

segregation due to the inclusion of rubber particles. Rubberised concrete mix was found

to be homogenous throughout all layers of the specimens.

3. Rubberised concrete significantly improved ductility and energy absorption of CFST and

CFDST by up to 2.5 times.

4. Existing Methods of for the predictions of ultimate peak compressive load for CFST and

CFDST filled with normal concrete yield good results for rubberised concrete columns

counterparts.

5. CFST and CFDST using cold-formed CHS significantly improved the ultimate peak

strength in RuC and normal concrete due to lateral confinement. Additional strength

through such confinement was achieved in all specimens.

6. Concrete steel bond zone at the inner and outer steel sections was seen to be promising

with both rubberised and normal concrete having ductile behaviour due to confinement.

7. With further research, RuCFDST has promising potential for applications such as

structural columns in seismic active zones.

Acknowledgments

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

23

The authors would like to deeply thank Liam O’keefe from Tyres Stewardship Australia and

Adrian Jones from Tyrecycle. Thanks are given to Andrew Sarkady and Anup Chakrabortty

from BASF for kindly donating the superplasticizer required for all the specimens. Thanks

are given the following technicians Matt Arpin, Malcolm Stafford, Jim Waters and Brad Rose

for assisting the students in performing the experiments. Thanks are given to Cameron

Marshall and Armin Hosseini, David Pegrum and Aarin Ryan, former students of UWA for

performing the tests and processing the test data.

REFERENCES

[1] K. B. Najim and M. R. Hall, "A Review of the Fresh/Hardened Properties and

Applications for Plain (PRC) and Self-Compacting Rubberised Concrete (SCRC),"

Construction and Building Materials, vol. 24, pp. 2043-2051, 2010.

[2] Hyder Consulting, “Stocks & fate of end of life tyres – 2013-14 study,” pp. 3, 2015.

[3] M. Elchalakani, "High Strength Rubberized Concrete Containing Silica Fume for the

Construction of Sustainable Road Side Barriers," in Structures, 2015, pp. 20-38.

[4] N. N. Eldin and A. B. Senouci, "Observations on Rubberized Concrete Behavior",

Cement, Concrete and Aggregates, vol. 15, pp. 74-84, 1993.

[5] M. Elchalakani, T. Aly, and E. Abou-Eisha, "Mechanical Properties of Rubberised

Concrete for Road Side Barriers", Australian Journal of Civil engineering, Volume

14, Issue 1, pages 1-12, 2016.

[6] T. Naka, B. Kato, and H. Kanatani, "Experimental Study on Welded Tubular

Connections," Research Institute of Welding, University of Tokyo, Japan, 1964.

[7] H.A. Toutanji, "Cement and Concrete Composites", 1996, Vol.18(2), pp.135-139

[Peer Reviewed Journal]

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

24

[8] N. I. Fattuhi and L. A. Clark, "Cement-Based Materials Containing Shredded Scrap

Truck Tyre Rubber", Construction and Building Materials, vol. 10, pp. 229-236,

1996/06/01 1996.

[9] N. Segre and I. Joekes, "Use of Tire Rubber Particles as Addition to Cement Paste",

Cement and Concrete Research, vol. 30, pp. 1421-1425, 2000.

[10] Z. K. Khatib and F. M. Bayomy, "Rubberized Portland Cement Concrete", Journal of

Materials in Civil Engineering, vol. 11, p. 206, 1999.

[11] S. Morino and K. Tsuda, "Design and Construction of Concrete-Filled Steel Tube

Column System in Japan", Earthquake Engineering and Engineering Seismology, vol.

4, pp. 51-73, 2003.

[12] O. Youssf, M. A. ElGawady, J. E. Mills, and X. Ma, "An Experimental Investigation

of Crumb Rubber Concrete Confined by Fibre Reinforced Polymer Tubes",

Construction and Building Materials, vol. 53, pp. 522-532, 2014.

[13] A. Duarte, B. Silva, N. Silvestre, J. de Brito, E. Júlio, and J. Castro, "Tests and

Design of Short Steel Tubes Filled with Rubberised Concrete", Engineering

Structures, vol. 112, pp. 274-286, 2016.

[14] Hassanein, M.F., Kharoob O.F., "Compressive Strength of Circular Concrete-filled

Double Skin Tubular Short Columns", Thin-Walled Structures, Vol. 77, pp. 165-173,

2014.

[15] Hassanein, M.F., Kharoob O.F., Liang, Q.Q., (2013), "Circular Concrete-Filled

Double Skin Tubular Short Columns with External Stainless Steel Tubes under Axial

Compression", Thin-Walled Structures, Vol. 73, pp. 252-263, 2013.

[16] Hassanein, M.F., Kharoob O.F., "Analysis of Circular Concrete-filled Double Skin

Tubular Slender Columns with External Stainless Steel Tubes ", Thin-Walled

Structures, Vol. 79, pp. 23-37, 2014.

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

25

[17] X.-L. Zhao, R. Grzebieta, and M. Elchalakani, "Tests of Concrete-Filled Double

Skin CHS Composite Stub Columns", Steel and Composite Structures, vol. 2, pp.

129-146, 2002.

[18] Z. Tao, L. Han, X. Zhao, "Behaviour of Concrete-Filled Double Skin (CHS inner

and CHS outer) Steel Tubular Stub Columns and Beam-Columns", Journal of

Constructional Steel Research, vol. 60, no. 8, pp. 1129-1158, 2004.

[19] BS. ENV 1994-1-1. Eurocode 4: design of composite steel and concrete structures –

Part 1-1: general rules and rules for buildings. UK: British Standards Institution

London; 1994.

[20] Australian/New Zealand Standards. Cold-formed structural steel hollow sections,

AS1163-2009. Sydney, Australia: Australia/New Zealand Standards; 2009.

[21] M. Elchalakani, "A Closed-Form Solution for Elastic Buckling of Thin-Walled

Unstiffened Circular Cylinders in Pure Flexure", Thin-Walled Structures, vol. 80, pp.

120-129, 2014.

[22] Australian/New Zealand Standards. General purpose and blended cements, AS3972-

2010. Sydney, Australia: Australia/New Zealand Standards; 2010

[23] Australian/New Zealand Standards. Test sieving procedures, AS3638-1993. Sydney,

Australia: Australia/New Zealand Standards; 1993

[24] G. Li, G. Garrick, J. Eggers, C. Abadie, M. A. Stubblefield, and S.-S. Pang, "Waste

tire Fiber Modified Concrete", Composites Part B: Engineering, vol. 35, pp. 305-312,

2004.

[25] E. Güneyisi, M. Gesoğlu, and T. Özturan, "Properties of Rubberized Concretes

Containing Silica Fume", Cement and Concrete Research, vol. 34, pp. 2309-2317,

2004.

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

26

[26] A. C. Ho, A. Turatsinze, R. Hameed, and D. C. Vu, "Effects of Rubber Aggregates

from Grinded Used Tyres on the Concrete Resistance to Cracking", Journal of

Cleaner Production, vol. 23, pp. 209-215, 2012.

[27] A. Turatsinze and M. Garros, "On the Modulus of Elasticity and Strain Capacity of

Self-Compacting Concrete Incorporating Rubber Aggregates", Resources,

conservation and recycling, vol. 52, pp. 1209-1215, 2008.

[28] O. Karahan, E. Özbay, K. Hossain, M. Lachemi, and C. D. Atiş, "Fresh, Mechanical,

Transport, and Durability Properties of Self-Consolidating Rubberized Concrete",

ACI Materials Journal, vol. 109, 2012.

[29] W. H. Yung, L. C. Yung, and L. H. Hua, "A study of the Durability Properties of

Waste Tire Rubber Applied to Self-Compacting Concrete", Construction and Building

Materials, vol. 41, pp. 665-672, 2013.

[30] O. Youssf, J. Mills and R. Hassanli, "Assessment of the Mechanical Performance of

Crumb Rubber Concrete", Construction and Building Materials, vol. 125, pp. 175-

183, 2016.

[31] M. Balaha, A. Badawy, and M. Hashish, "Effect of Using Ground Waste Tire

Rubber as Fine Aggregate on the Behaviour of Concrete Mixes", Indian Journal of

Engineering and Materials Sciences, vol. 14, p. 427, 2007.

[32] I. Mohammadi, H. Khabbaz and K. Vessalas, "Enhancing Mechanical Performance

of Rubberised Concrete Pavements with Sodium Hydroxide Treatment", Materials

and Structures, vol. 49, no. 3, pp. 813-827, 2015.

[33] AS 1012.9, Methods of testing concrete - Compressive strength tests - Concrete,

mortar and grout specimens, Standards Australia, 2014.

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

27

[34] Abendeh, R., Ahmad, H.S., Hunaiti, Y.M., "Experimental studies on the behavior of

concrete-filled steel tubes incorporating crumb rubber", Journal of Constructional

Steel Research, Vol. 122, pp. 251-260, 2016.

[35] Duarte, A.P.C., Silva, B.A., Silvestre, N., de Brito, J., Júlio, E., Castro, J.M., "Finite

element modelling of short steel tubes filled with rubberized concrete", Composite

Structures, Vol. 150, pp. 28-40, 2016.

[36] Zhao, X.L., Han L.H. "Double skin composite construction", Prog Struct Mat Eng,

Vol. 8(3), pp. 93-102, 2006.

[37] Yu, M., Zha, X., Ye, J., Li, Y., "A unified Formulation for Circle and Polygon

Concrete-Filled Steel Tube Columns Under Axial Compression", Engineering

Structures, Vol. 49, pp. 1-10, 2013.

[38] Australian/New Zealand Standards. Steel Structures, AS4100-1998. Sydney,

Australia: Australia/New Zealand Standards; 1998.

[39] M. Pagoulatou, T. Sheehan, X. Dai and D. Lam, "Finite Element Analysis on the

Capacity of Circular Concrete-Filled Double-Skin Steel Tubular (CFDST) Stub

Columns", Engineering Structures, vol. 72, pp. 102-112, 2014.

[40] Elchalakani, M., Zhao, X. L. and Grzebieta, R. H. (2002), "Tests on Concrete-filled

Double-Skin Composite Short Columns under Axial Compression", Thin-Walled

Structures, Vol. 40, No. 5, pp. 415-441, Elsevier, UK.

[41] Zhao, X.L., Grzebieta R.H. and Elchalakani, M. (2002), "Tests of Concrete-filled

Double-Skin CHS Stub Columns", Steel and Composite Structures, Vol. 2, No. 2, pp.

129-146, 2002.

[42] Zhao, X.L., Grzebieta R.H., Uker A., and Elchalakani, M. (2002), "Tests of

Concrete-filled Double-Skin SHS outer and CHS inner-Composite Stub Columns",

Advances in Steel Structures, An International Journal, Vol. 1, No. 2, pp. 567-573..

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

28

Notation

RuC Rubberised concrete

RA Rubber Aggregate

CA Course Aggregate

FA Fine Aggregate

NaOH Sodium Hydroxide

SCC Self-compacting concrete

CFST Concrete-filled steel tube

CFDST Concrete-filled double-skin steel tube

RuCFST Rubberised concrete-filled steel tube

RuCFDST Rubberised concrete-filled double-skin steel tube

FRP Fibre reinforced polymer

CHS Circular hollow section

SHS Square hollow section

EC4 Eurocode 4

PSD Particle size distribution

DI Ductility Index

LVDT Linear variable differential transformer

D Diameter of steel tube

Do Diameter of outer steel tube

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

29

Di Diameter of inner steel tube

Dc Diameter of filled concrete

tso Thickness of outer steel tube

tsi Thickness of inner steel tube

L Length of test specimen

λso Circular hollow section slenderness

As Cross-sectional area of steel

Ac Cross-sectional area of concrete

Ac,nominal Nominal cross-sectional area of concrete

σyt Yield stress of steel

fsyo Yield strength of outer steel tube

fsyi Yield strength of inner steel tube

f’c Concrete compressive strength

fck Characteristic concrete strength

Pu Experimental ultimate peak axial compressive load

Pyn Predicted nominal empty hollow section ultimate peak axial compressive load

Pyt Predicted empty hollow section ultimate peak axial compressive load

Pul,Zh Zhao’s predicted composite ultimate peak axial compressive load

Pul,Has Hassanein’s predicted composite ultimate peak axial compressive load

Pul,Tao Tao’s predicted composite ultimate peak axial compressive load

Pul,EC4 Eurocode 4 predicted composite ultimate peak axial compressive load

Pcr Critical buckling load

χ Hollow section ratio

ξ Enhancement factor due to confinement effects (Tao)

nao Reduction factor for outer steel (EC4)

nco Enhancement factor due to confinement effects (EC4)

Relative slenderness of specimen

(EI)eff Effective flexural stiffness of the specimen cross-section

Fig. 1: Specimen details of a RuCCFDST

Fig. 2: Rubber aggregate size

Figure

Fig. 3: Particle size distribution plot of aggregate

Fig. 4: Aggregate mix distribution

Fig. 5: Specimen CHS-O114-42-00 prior to testing on the DLS500 (strain gauges not shown)

Fig. 6: Load-deflection curve of (a) Series 1 Empty CHS and (b) Series 2 Empty CHS

(b)

(a)

Fig. 7: Buckling mode of a family of ten CHS specimens

Fig. 8: Load-deflection curves of CHS-O165 specimen for 0%, 15% and 30% rubber aggregate replacement

Fig. 9: CHS-O165-I42 (C5 outer and C1 inner) specimen load-deflection curve

Fig. 10: CHS-O165-I89 (C5 outer and C2 inner) specimen load-deflection curve

Fig. 11: CHS-O114-3.2-I42 (C4 outer and C2 inner) specimen load-deflection curve

Fig. 12: CHS-O114-3.6-I42 (C4 outer and C3 inner) specimen load-deflection curve

Fig. 13: Experimental progressive axial loading of specimen CHS-O165-89-30 (cont.)

Fig. 13: Experimental progressive axial loading of specimen CHS-O165-89-30

Fig. 14: Specimen CHS-O165-89-30 internal buckling of the inner CHS

Fig. 15: CHS-O165-89-30 buckle formation photos corresponding to location within load-

deflection curve

Fig. 16: Concrete ductile filling at the buckled steel areas (the percentage of rubber replacement is indicated at the bottom right of each photograph)

Fig. 17: Rubberised concrete post-failure segregation

Fig. 19: Load deflection curve of CHS-O165-I89-30

Fig. 18: CHS-O165-I89 measured axial load verses strain for 0%, 15% and 30%

Fig. 20: Ductility index of all test specimens

Table 1: Measured properties of the CHSs

Specimen No.

Diameter (D) [mm]

Thickness (t) [mm]

Area (As) [mm2] D/t

Yield stress [MPa]

Squash load [kN]

C1 42.4 2.6 325 16.31 33.3 510 166 C2 88.9 3.2 862 27.78 54.5 490 422 C3 114.3 3.2 1117 35.72 59.3 415 464 C4 114.3 3.6 1252 31.75 56.5 445 557 C5 165.1 3.5 1777 47.17 74.5 395 702

Table 2: Chemical Composition of Cement (w %)

Table 3: Mass and Density of Concrete Compressive Cylinders

Specimen Name Density (kg/m3) Average Density (kg/m3)

Concrete Cylinder Strength (MPa)

Average Cylinder Strength (MPa)

CT-00-01 2256.8 2271

47.4 50.27 CT-00-02 2279.1 52.2

CT-00-03 2275.9 51.2 CT-15-01 2088.1

2086 24.9

24.95 CT-15-02 2084.9 21.1 CT-15-03 2084.9 25.0 CT-30-01 1948.1

1943 13.7

14.37 CT-30-02 1941.7 14.4 CT-30-03 1938.5 15.0

Table 4: Measured properties of CFST and CFDST specimens

Specimen No. Specimen CHS composition (Table 1)

Area of concrete (Ac) [mm2] (As/Ac)%

CHS-O165 C5 19631 9.05 CHS-O114-3.2-I42 C1, C3 7732 18.65 CHS-O114-3.6-I42 C1, C4 7597 20.76

CHS-O165-I42 C1, C5 18220 11.54

Cement Type SiO2 CaO Al2O3 Fe2O3 MgO SO3 LOI Na2O Swan Grey Cement

Type GP (%) 20.6 63.5 5.2 3.0 1.3 2.6 1.8 0.5

Table

CHS-O165-I89 C2, C5 13424 19.65

Table 5: Summary of ultimate peak loads (average of 2 specimens)

Specimen No. Pu,1 (Trial 1) kN Pu,2 (Trial 2) kN Pu,avg kN

C1 171 170 170.5 C2 443 443 443 C3 430 436 433 C4 559 564 561.5 C5 663 702 682.5

Table 6: Peak experimental loads for CFDST/CFST specimens

Specimen No. Rubber Content

0% 15% 30% Force (kN) Force (kN) Force (kN)

CHS-O165 1876 1291 (31.2%) 1130 (39.8%) CHS-O114-3.2-I42 1073 824 (23.2%) 742 (30.8%) CHS-O114-3.6-I42 1255 968 (22.9%) 900 (28.3%)

CHS-O165-I42 1888 1428 (24.4%) 1245 (34.1%) CHS-O165-I89 1855 1561 (15.8%) 1441 (22.3%)

Table 7: Comparison between experimental strengths with different design model strengths

Specimen [kN]

[kN]

[kN]

[kN]

[kN]

CHS-O165-0 1876 1541 1603 1574 1582 0.82 0.85 0.84 0.84 CHS-O165-15 1291 1118 1189 1241 1215 0.87 0.92 0.96 0.94 CHS-O165-30 1130 942 1016 1101 1063 0.83 0.90 0.97 0.94

CHS-O114-3.2-I42-00 1073 960 1019 1007 1101 0.89 0.95 0.94 1.03 CHS-O114-3.2-I42-15 824 793 851 876 956 0.96 1.03 1.06 1.16 CHS-O114-3.2-I42-30 742 724 781 821 897 0.98 1.05 1.11 1.21 CHS-O114-3.6-I42-00 1255 1048 1131 1120 1240 0.83 0.90 0.89 0.99 CHS-O114-3.6-I42-15 968 884 964 991 1096 0.91 1.00 1.02 1.13 CHS-O114-3.6-I42-30 900 816 894 937 1039 0.91 0.99 1.04 1.15

CHS-O165-I42-00 1888 1646 1704 1677 1658 0.87 0.90 0.89 0.88 CHS-O165-I42-15 1428 1254 1319 1368 1325 0.88 0.92 0.96 0.93 CHS-O165-I42-30 1245 1090 1159 1239 1187 0.88 0.93 1.00 0.95 CHS-O165-I89-00 1855 1698 1740 1720 1610 0.92 0.94 0.93 0.87 CHS-O165-I89-15 1561 1409 1457 1492 1374 0.90 0.93 0.96 0.88 CHS-O165-I89-30 1441 1288 1339 1397 1278 0.89 0.93 0.97 0.89

Average 0.89 0.94 0.97 0.99 Standard deviation (SD) 0.04 0.05 0.07 0.12