Influence of concrete mix proportions on axial performance of concrete-filled steel tubes made with self-compacting concrete

Abstract

Self-compacting concrete is quite commonly used in concrete-filled steel tube structures, but the compaction level of the self-compacting concrete, that is, the percentage of volume occupied by materials other than air void, within the steel tube is seldom investigated. The authors are of the view that the concrete mix proportions of the self-compacting concrete may have significant effects on the compaction level of the self-compacting concrete, which will be quantified by the ‘compaction index’ proposed in this study and thus the performance of the concrete-filled steel tube. Moreover, the mix proportions would also influence the performance of the concrete-filled steel tube by affecting the aggregate–aggregate and aggregate–paste interactions of the concrete, albeit this important issue is rarely addressed in previous studies either. Herein, a pilot study is conducted to investigate the influences of the self-compacting concrete mix proportions on the axial performance of concrete-filled steel tube. Four groups of concrete-filled steel tube specimens made with different self-compacting concrete were tested, and the investigated concrete mix parameters included the paste volume, fine to coarse aggregate ratio, and 9.5–19.0 mm aggregate ratio. It was found that the compaction index of the self-compacting concrete is a key factor enabling the successful use of self-compacting concrete in concrete-filled steel tube. Moreover, the paste volume and aggregate proportions of the concrete mix have certain effects on the post-peak behaviour and ductility of concrete-filled steel tube.

Keywords

compaction index concrete-filled steel tube concrete mix proportions self-compacting concrete

Introduction

In the last two decades, concrete-filled steel tube (CFST) structures have gradually gained popularity in high-rise buildings (Han et al., 2014) and long-span bridges (Chen, 2008), owning to their various advantages in structural performance and constructability. These advantages include but not limited to (1) enhancing the strength and ductility of the concrete infill through confinement by the steel tube, (2) alleviating premature local buckling of the steel tube with the concrete infill providing lateral restraint and (3) improving the constructability because the steel tube can serve as a permanent formwork. Although the steel tubes being at the exterior would be susceptible to corrosion and fire attacks, these problems have been resolved by providing corrosion/fire-resistant coatings or encasement.

In CFST structures, concrete placement of good quality has always been demanded to achieve high structural performance. However, satisfactory concrete placement is hard to achieve because the access for any means of mechanical vibration to compact the fresh concrete inside the steel tube is usually rather limited. Self-compacting concrete (SCC, also known as self-consolidated concrete) is an effective solution in such circumstances (Okamura and Ouchi, 2003). A good example is the construction of Bosiden Yangtze River Bridge in Sichuan, China, which was completed in 2012 (Wang et al., 2015). This bridge comprised several CFST arch ribs with 530 m span, and the SCC infill of each arch rib was pumped from the bottom and sucked to the top. Without any viable means of internal vibration, the placement quality of the concrete solely relied on its self-compactability. In the end, a concrete mix with 28-day cube strength above 80 MPa and a flow spread larger than 600 mm was adopted for the bridge.

It is worth noting that back in 2004, Han and Yao (2004) already conducted 38 compression load tests of SCC-filled steel tube columns to study the influence of different compaction methods on the structural performances of CFST columns. The test parameters in the experimental programme included the method of compaction (by hand, by poker vibrator and self-compacting), loading configuration (axial load with or without eccentricity), cross-sectional shape (circular or square) and slenderness ratio. The general conclusion was that the specimens made by the different compaction methods had similar basic properties, but those made with concrete compacted by poker vibrator obviously had higher loading capacities.

As a matter of fact, ever since Okamura and Ouchi (2003) developed the SCC in 2003, attention has been drawn to the possible application of SCC in CFST structures. For instance, Han et al. (2005a) developed models for the axial load capacities and axial load–deformation relations of stub columns with circular or square sections under compression. Han et al. (2005b) reported cyclic load tests of high-strength and ultra-high-strength SCC-filled steel tube columns, which revealed that the cyclic behaviours of CFST with SCC infill were similar to those with normal concrete infill. Han et al. (2006) explored the flexural behaviours of SCC-filled steel tube beams and found that the beams filled with SCC behaved similarly to those filled with normal concrete. Yu et al. (2007) conducted axial compression tests on SCC-filled steel tube stub columns of which some had holes or slots on the steel tubes and some had partial contact at both loading ends, to study their mechanical behaviours under special conditions. Yu et al. (2008) performed compression tests on stub column and beam-column CFSTs filled with ultra-high-strength SCC with cube strength up to 121.6 MPa and demonstrated that the CFSTs had good ultimate strength and ductility.

Later, Lu et al. (2009, 2010a, 2010b) tested the fire performance of SCC-filled steel tube structures. At the same time, Zhao and Packer (2009) expanded the territory of this research field by conducting an experimental study on axially loaded SCC-filled steel tube columns with elliptical sections. Zhu et al. (2010) studied the effects of inner steel I sections and cruciform sections on the behaviour of SCC-filled tubular columns and found that the failure modes of the specimens with inner steel sections were quite different. In 2010–2012, Han and Li (2010) and Li and Han (2011, 2012) conducted a series of investigations to study the seismic behaviours of joints connecting SCC-filled steel tube columns and steel beams. More recently, Xiong et al. (2017a, 2017b) conducted a series of tests on high-strength SCC-filled steel tube columns and revealed that the high-strength SCC-filled steel tube columns were quite ductile and the coarse aggregate ratio has little effect on the performance of the columnar specimens. On the other hand, the shape of the cross section is another research focus in this field, as covered by a series of studies on the behaviours of CFST columns with cross sections of unconventional shapes (Han et al., 2010; Liu et al., 2017; Ren et al., 2014; Xu et al., 2018; Zhu et al., 2019).

Although many structural aspects of CFST structures with SCC infill have been investigated in previous studies, it is noticed that the mix proportions of SCC are quite different from one study to another, and the possible influences of the concrete mix proportions of the SCC on the performance of CFST members have not been explored yet. In this study, four concrete mix proportions with the same water/cementitious materials (W/CM) ratio but different paste volumes, different combinations of fine and coarse aggregate and different mixes of coarse aggregates were adopted. The flowability, cohesiveness and passing ability of each mix were measured. Besides, a new measure designated as ‘compaction index’, which is based on a common practice to measure the wet density of the fresh concrete mix (Li and Kwan, 2015; Ling and Kwan, 2018; Maybury et al., 2017; Wong and Kwan, 2008), was proposed to quantify the compaction level, that is, the percentage of volume occupied by materials other than air void, of each mix inside the steel tube.

Four groups of circular-sectioned CFST specimens corresponding to the four SCC mixes were prepared. Since SCC is indeed a type of high-performance concrete with better rheological properties, thus easier to compact, the use of SCC together with certain means compaction would ensure the placing quality of concrete for CFST structures. Therefore, for each group, one set of the CFST specimens were cast with self-compaction only, that is, under the uncompacted condition, and the other set were cast with the aid of poker vibration, that is, under the compacted condition, in order to identify the variation in compaction index of the corresponding SCC mix. The CFST specimens were tested under axial compression, and the full-range axial performance of the CFST was correlated to the mix parameters and properties of the SCC. The authors are of the view that the mix proportions and compaction index of the SCC could have major effects on the axial performance of CFST made with SCC. In the actual construction, when the accessibility for compaction becomes the control factor, the findings in this study can also provide suggestions on the choice of mix proportions of SCC.

Experimental programme

Mix proportions and properties of concrete

The concrete mixes designed for this study comprised cementitious materials, water, superplasticizer, coarse aggregate and fine aggregate, whose densities were measured and listed in Table 1. Meanwhile, the particle size distributions of the solids (i.e. cementitious materials, coarse aggregate and fine aggregate) are presented in Figure 1. The cementitious materials included Grade PO 42.5 ordinary Portland cement (OPC) and condensed silica fume (CSF) complying with the Chinese National Standards GB 175-2007 (2007) and GB/T 27690-2011 (2011), respectively. For the OPC, more than 75% has particle size falling within the range of 3.0–20.0 μm. For the CSF, more than 75% has particle size smaller than 1.0 μm. Crushed granite was used as the coarse aggregate, which has two nominal maximum particle sizes, namely 9.5 and 19.0 mm. River sand with a nominal maximum particle size of 1.18 mm was used as the fine aggregate. The superplasticizer added to each concrete mix was a polycarboxylate-based chemical admixture having a solid content of 20%.

Table 1.

Densities of concrete ingredients.

Material	Density (kg/m³)
Water	1000
OPC (grade P.O. 42.5)	3112
CSF	2200
Superplasticizer	1030
Fine aggregate	2652
9.5 mm coarse aggregate	2683
19.0 mm coarse aggregate	2683

OPC: ordinary Portland cement; CSF: condensed silica fume.

Four types of concrete mixes with two distinct values of paste volume, defined as the volume of paste (the paste comprises the cementitious materials, water and superplasticizer) expressed as a percentage of the volume of concrete, were adopted for the experimental programme, and their mix proportions are listed in Table 2. For all the concrete mixes, the CSF content by weight of the total cementitious materials was 15% and the W/CM ratio was 0.41. For the Type 1 mix, the paste volume was 40.8%, whereas for the Type 2 series (i.e. Type 2A, Type 2B and Type 2C), the paste volume in each mix was 36.7%. The Type 1 and Type 2A mixes have a fine aggregate to coarse aggregate ratio of 1:1 and a 9.5 mm coarse aggregate to 19.0 mm coarse aggregate ratio of 1:1. In contrast, the Type 2B and Type 2C mixes have different aggregate proportions; the Type 2B mix has a fine aggregate to coarse aggregate ratio of 1:1 and a 9.5 mm coarse aggregate to 19.0 mm coarse aggregate ratio of 1:0, whereas the Type 2C mix has a fine aggregate to coarse aggregate ratio of 1.5:1 and a 9.5 mm coarse aggregate to 19.0 mm coarse aggregate ratio of 1:1.

Table 2.

Concrete mix proportions.

Concrete type	Type 1	Type 2A	Type 2B	Type 2C
Corresponding CFST specimens	T1-S1 T1-S2 T1-V1 T1-V2	T2A-S1 T2A-S2 T2A-V1 T2A-V2	T2B-S1 T2B-S2 T2B-V1 T2B-V2	T2C-S1 T2C-S2 T2C-V1 T2C-V2
OPC (kg/m³)	459	408	408	408
CSF (kg/m³)	80	71	71	71
Superplasticizer (L/m³)	10	10	10	10
Water (kg/m³)	215	191	191	191
Fine aggregate (kg/m³)	790	840	840	1008
9.5 mm aggregate (kg/m³)	395	420	840	336
19.0 mm aggregate (kg/m³)	395	420	0	336
W/CM ratio	0.41	0.41	0.41	0.41
Paste volume (%)	40.8	36.7	36.7	36.7

CFST: concrete-filled steel tube; OPC: ordinary Portland cement; CSF: condensed silica fume; W/CM: water/cementitious materials.

The fresh properties of each concrete mix were measured by the slump-flow test (BS EN 12350-2, 2009; BS EN 12350-8, 2010), U-box test (Okamura and Ouchi, 2003) and sieve segregation test (BS EN 12350-11, 2010) of which the results are listed in Table 3. The flow spread results obtained from the slump-flow tests are measures of flowability; the U-box filling height results from the U-box tests are measures of passing ability and the sieve segregation index (SSI) results from the sieve segregation tests are measures of cohesiveness (note that a higher SSI implies lower cohesiveness and a lower SSI implies higher cohesiveness). Based on these test results, all the concrete mixes, that is, the Type 1, Type 2A, Type 2B and Type 2C mixes, were found to have reasonably high flowability, passing ability and cohesiveness. Among them, the Type 2B mix has the best flowability and passing ability, and the Type 1 and Type 2A mixes have relatively lower flowability and passing ability.

Table 3.

Concrete properties.

Concrete type	Type 1	Type 2A	Type 2B	Type 2C
Slump (mm)	250	252	249	251
Flow spread (mm)	510	500	561	524
U-box filling height (mm)	271	266	311	285
Sieve segregation index (%)	1.40	1.33	1.58	0.81
Age at strength test (days)	86	86	72	72
Cylinder strength (MPa)	81.0	74.4	69.6	54.7
Young’s modulus (GPa)	30.4	31.1	30.2	25.7

From each concrete mix, four cylinders with 150 mm diameter and 300 mm height were made to determine the cylinder compressive strength of the concrete. During casting, all the concrete cylinders were not subjected to any vibration process, that is, with no compaction applied, and their consolidation relied solely on their self-compactability. During the compression strength test, the axial displacement was applied at a constant rate of 0.3 mm/min. The second last row of Table 3 presents the average compressive strength of the four cylinders cast from each concrete mix. From these results, it can be seen that the four different concrete mixes have different compressive strengths, and the order from the highest compressive strength of 81.0 MPa to the lowest compressive strength of 54.7 MPa is Type 1, Type 2A, Type 2B and Type 2C. It is noteworthy that the four concrete mixes actually have the same W/CM ratio. So, it should be the differences in paste volume and aggregate proportions that caused the difference in uncompacted strength.

During the compressive tests of the cylinders, Young’s modulus of the concrete was also measured. The last row of Table 3 presents Young’s modulus of each type of concrete tested. It is seen that Young’s moduli of the Type 1, Type 2A and Type 2B mixes were similar, all being slightly above 30.0 GPa. However, the Type 2C mix, which has the lowest compressive strength of 54.7 MPa, also has the lowest Young’s modulus of 25.7 GPa. Hence, although the four concrete mixes have the same W/CM ratio, their Young’s moduli were not the same. It is evident that the mix parameters, which affected the compressive strength, should have also affected Young’s modulus.

Properties of steel tubes

Hot-rolled seamless steel tubes were used in the experimental programme. Only one type of section, circular section, with a nominal outer diameter of 194 mm and a nominal thickness of 6 mm was adopted. Six longitudinal coupons were cut from two of the steel tubes for tensile tests as per BS 18-87 (1987), which then gave the mean values of the mechanical properties of the steel as follows: yield strength = 342 MPa, Young’s modulus = 212 GPa, ultimate strength = 503 MPa and Poisson’s ratio = 0.29. During the coupon tests, it was found that the yield plateaus of the steel normally ended at an axial strain of 2.5%–3.0%, which is useful information for the analysis of the axial stress–strain behaviour of the CFST specimens later.

CFST specimens

Four groups of CFST short column specimens were fabricated, each group consisting of four specimens exclusively made with one of the four types of concrete mixes, as depicted in the first two rows of Table 2. In each group of CFST specimens, two of the specimens were cast of the concrete mix without any compaction applied (i.e. uncompacted just like placing of SCC) and the other two specimens were cast of the concrete mix with internal vibration applied using a poker vibrator (i.e. compacted just like placing of ordinary concrete). Such arrangement was to assess the influence of the concrete mix proportions on the compressive behaviours of the CFST columns under the two different compaction conditions. Moreover, each test setting consisted of two repeated tests to crosscheck the test results and increase the confidence level of the test findings. The notations of the CFST specimens have the following meanings: ‘T1’, ‘T2A’, ‘T2B’ and ‘T2C’ stand for the type of concrete mix used for the concrete infill; ‘S’ means casting of concrete infill solely by self-compaction (no compaction applied); ‘V’ means casting of concrete infill with vibration applied (compaction applied) and the number following ‘S’ or ‘V’ denotes specimen 1 or specimen 2 of the two repeated specimens.

Compaction index of concrete

In most previous research on SCC, the focus was mainly on the flowability required for the concrete mix to be qualified as an SCC, and a flow spread of 550 mm (Han et al., 2006; Lu et al., 2009, 2010a, 2010b; Zhu et al., 2010) was generally regarded as the minimum for attaining self-compactability, that is, for the concrete mix to be considered suitable for placing without compaction. However, the performance of the concrete with no compaction applied was seldom compared to that of the concrete with compaction applied to verify the suitability of the concrete mix to be placed without compaction. In this regard, the European Standard BS EN 12350-4 (2009) has stipulated a test method of measuring the ‘degree of compactability’ of a fresh concrete mix as the uncompacted volume to compacted volume ratio (same as the compacted density to uncompacted density ratio). However, the authors are of the view that even under the compacted condition, there could still be entrapment of air inside the concrete mix, hence both the compaction level under uncompacted condition and the compaction level under compacted condition should be measured for evaluation, using the test method developed in the following.

While filling the fresh concrete into the steel tube, as shown in Figure 2, the opportunity was taken to determine the apparent wet density of the concrete so as to evaluate the compaction level of the concrete. The apparent wet density of the concrete was determined by measuring the mass of concrete $M_{c}$ in the steel tube and the volume $V_{c}$ inside the steel tube and calculating the apparent wet density as $M_{c} / V_{c}$ . However, given the mix proportions of the concrete mixture, the theoretical wet density $ρ_{con}$ may be calculated as follows

ρ_{c o n} = ρ_{α} R_{α} + ρ_{β} R_{β} + ρ_{γ} R_{γ} + \dots

(1)

in which ρ and R stand for the density of a specific material and the volumetric ratio of the specific material to all the materials in the mixture, respectively; each of the subscripts α, β and γ stands for each material, and for this study, the various materials are the water, OPC, CSF, superplasticizer, coarse aggregate and fine aggregate. The density ρ of each material in the mixture can be obtained from Table 1, whereas the volumetric ratio of each material in the material can be calculated from the mix proportions given in Table 2. This formula for calculating the overall density of a mixture has been used by Wong and Kwan (2008) in their so-called wet packing method to calculate the solid concentration and packing density of a mixture of particles. It is proposed herein to adopt the apparent wet density to theoretical wet density ratio, that is, $(M_{c} / V_{c}) / ρ_{con}$ , as the ‘compaction index’ to quantify the compaction level of the concrete mixture in the steel tube. Theoretically, due to the possible entrapment of air inside the concrete mixture, the compaction index should be lower than or at most equal to 1.0, but in fact some of the specimens under the compacted condition had compaction indices marginally greater than 1.0 due to experimental errors.

The compaction indices of the fresh concrete mixes filled into the CFST specimens are presented in the form of a bar chart in Figure 3. Among the specimens T1-S, T2A-S, T2B-S and T2C-S with the concrete filled by self-compaction (no compaction applied), T1-S has the highest compaction index of 1.002. Hence, under the uncompacted condition, the Type 1 mix has the best compaction index. The rather high compaction index of the Type 1 concrete mix under the uncompacted condition verifies that this concrete mix is well qualified as an SCC suitable for placing with no compaction applied, albeit its flow spread is only 510 mm (i.e. smaller than 550 mm which is the traditional requirement of flow spread for SCC) and U-box filling height is only 271 mm (i.e. smaller than 300 mm which is the common target of U-box filling height for SCC). Somehow, the compaction indices of the four concrete mixes under the uncompacted condition do not follow the order of flow spread value, indicating that a high flow spread value does not necessarily guarantee a high compaction index. Since the Type 1 mix has the largest paste volume, it seems that the paste volume is a more important factor governing the compaction index of the concrete mix.

However, among the specimens T1-V, T2A-V, T2B-V and T2C-V with the concrete filled and compacted by vibration (compaction applied), T1-V has the highest compaction index, and the descending order of compaction index of the concrete mixes follows the sequence of Type 1 > Type 2A > Type 2B > Type 2C. As expected, the compaction index of each concrete mix is higher under the compacted condition than under the uncompacted condition. More importantly, despite the application of compaction, the Type 2B mix has a compaction index of 0.992 and the Type 2C mix has a compactability of only 0.953, indicating that even with vibration applied, there could be 1% – 0.992 = 0.8% air voids in the Type 2B mix and about 1% – 0.953 = 4.7% air voids in the Type 2C mix. Hence, if the concrete mix does not have a good compactability under the compacted condition, there could be air entrapped in the concrete mix even with compaction applied. Since Type 2C mix has the highest fine to total aggregate ratio of 0.6 among the four concrete mixes, it seems that the aggregate proportion has certain effect on the compactability.

Traditionally, Type 2B would be the only SCC among the four mixes in this study according to the values of flow spread and U-box filling height. However, Figure 3 shows that both T1 and T2B have compacted compaction index close to 1.00 and uncompacted compaction index greater than 0.95, meaning such mix designs are able to achieve full compaction via means of vibration, while their self-compaction will lead to less than 5% void in volume. Therefore, it is reasonable to conclude that Type 1 and Type 2A can also achieve satisfactory self-compactability.

Instrumentation in compressive test of CFST specimens

Figure 4 shows the instrumentation for the axial compression test of the CFST specimens. Two vertical linear variable displacement transducers (LVDTs) were fixed diagonally to measure the average axial strain over the full height of the specimen. At every 90^o along the circumference of the mid-height section, there was a horizontal LVDT pointing towards the centre of the CFST section, a vertical strain gauge and a horizontal strain gauge. The contact points of the horizontal LVDTs at the CFST coincided with the positions of the vertical strain gauges, whereas the position of each horizontal strain gauge was offset to that of the nearest vertical strain gauge by 10 mm. The readings from the LVDTs and strain gauges, as well as the axial force from the testing machine, were recorded by a data logger.

The data from the horizontal LVDT were used to calculate the lateral strains at the mid-height section along two radial directions perpendicular to each other, while the readings from the horizontal strain gauges gave the circumferential laterals strains. Theoretically, the two types of lateral strains should be the same. Both the LVDT and strain gauge data can be used to calculate the lateral-to-axial strain ratio at the mid-height section of each specimen, but both have pros and cons at different stages of loading, as shown in Figure 5. In the initial stage, the readings for the strain gauges were more accurate, whereas the intrinsic reading errors in the LVDTs distorted the strain results when the displacement readings were small. However, in the later stage, as the deformation grew, the strain gauge readings were interfered by local buckling of the steel tube, and some of the strain gauges even stopped functioning at the very late stage due to their detachment from the steel tube. Meanwhile, as the errors were ‘diluted’ in larger displacement readings, the LVDT data became more reliable at the later stage of the experiment. Therefore, in the final results of lateral-to-axial strain ratio presented for each CFST specimen, the values have been consolidated in such a fashion that initially, the strain gauge results reigned, and once drastic fluctuation occurred, the LVDT results took over.

The displacement-controlled method was adopted for the axial compression tests of the CFST specimens, with a compressive displacement rate of 0.5 mm/min applied to the CFST specimens to match the strain rate in the compressive strength test of the concrete cylinders.

Axial compression test results of CFST specimens

The curves for the relations of the axial load versus axial strain and the lateral-to-axial strain ratio versus axial strain of the CFST specimens from the T1, T2A, T2B and T2C groups are plotted in Figures 6 to 9, respectively. Each figure shows the results of the four CFST specimens from the same group filled with the same type of concrete.

For all the CFST specimens, during the initial elastic stage, the values of the lateral-to-axial strain ratio were generally around 0.30, which is consistent with Poisson’s ratio of 0.29 of the steel coupons measured in this study. Given that Poisson’s ratio of concrete is normally between 0.15 and 0.20 (Fam and Rizkalla, 2001; Martínez-Lage et al., 2012; Yi et al., 2012), it is reasonable to postulate that there was delamination between the steel tube and the concrete infill (Ouyang et al., 2017; Ouyang and Kwan, 2018). At an axial strain of higher than 0.5%, the value of the lateral-to-axial strain ratio started to increase substantially. This means the effective lateral-to-axial strain ratio of concrete was then larger than that of the steel, owing to formation of splitting cracks and lateral dilation within the concrete, which caused the concrete to push the steel tube outwards. As a result, the passive confinement received by the concrete infill from the steel tube started to grow. Finally, the value of the lateral-to-axial strain ratio plateaued at a level between 0.65 and 1.00 (but mostly around 0.85) with the exact value varying from specimen to specimen.

The axial load–strain curves of all the CFST specimens have similar patterns, which can be summarized as follows. (1) As the axial strain increased, the axial load first increased linearly at a fairly rapid rate and then increased nonlinearly at a decreasing rate to a certain peak value. (2) After reaching the peak value, the axial load dropped to a certain minimum value at the trough of the curve and then bounced back with a gradually increasing trail up to an axial strain of at least 0.06. From each axial load–strain curve, three key axial load values may be obtained as indicators to reveal the differences in axial performance between the CFST specimens tested; these are the axial load at the peak (denoted by P_peak), the post-peak axial load at the trough (denoted by P_trough) and the residual strength at an axial strain of 0.06 (denoted by P_0.06), which are graphically presented in Figures 10 to 12, respectively.

For the T1 group of CFST specimens filled with Type 1 concrete, it can be seen from Figure 10 that the P_peak values of T1-V1 and T1-V2 are only marginally higher than those of T1-S1 and T1-S2, indicating that the non-application or application of compaction to the concrete infill had caused little change in the axial strength of the CFST specimen. This was because of the very high compaction index of the Type 1 concrete, which rendered the concrete self-compacting. Even without compaction applied to the concrete infill, T1-S1 and T1-S2 attained axial strengths of almost 3500 kN, which are the highest among the CFST specimens with no compaction applied to the concrete infill. Likewise, from Figures 11 and 12, it can also be seen that the compaction applied to the concrete infill had only marginally increased the P_trough and P_0.06 values of the T1 group of CFST specimens.

For the T2A group of CFST specimens filled with Type 2A concrete, it is noted from Figures 10 to 12 that the average values of P_peak, P_trough and P_0.06 of the T2A-S set are only 93.1%, 91.2% and 95.5%, respectively, of those of the T2A-V set. Hence, for this group of specimens filled with Type 2A concrete, the application of compaction had significantly increased the concrete strength and axial performance of the CFST and the non-application of compaction had caused certain detrimental effects. This was because the Type 2A concrete had a compaction index of 0.984 under the uncompacted condition and a compaction index of 1.010 under the compacted condition. Comparing the P_trough values of the T2A-V set to those of the T1-V set, it is interesting to note that the P_trough values of the T2A-V set are about 7.5% higher. In fact, from the axial load–strain curves of the T1-V set and the T2A-V set plotted in Figures 6 and 7, respectively, it can be seen that the T2A-V set has higher residual strengths within the post-peak range and thus higher ductility than the T1-V set.

For the T2B group of CFST specimens filled with Type 2B concrete, Figures 10 to 12 reveal that the average values of P_peak, P_trough and P_0.06 of the T2B-S set are only 86.2%, 89.3% and 92.2%, respectively, of their counterparts of the T2B-V set. This situation of lower axial performance due to non-application of compaction to the concrete infill seems to be worse than that of the T2A group of CFST specimens filled with Type 2A concrete. This was because the compaction index under the uncompacted condition of the Type 2B concrete was lower than that of the Type 2A concrete, albeit the Type 2B concrete had a higher flowability. Comparing the P_trough values of the T2B-V set to those of the T1-V set and the T2A-V set, it is noted that the P_trough values of the T2B-V set are similar to those of the T2A-V set and about 6.2% higher than those of the T1-V set. This phenomenon can also be seen from the axial load–strain curves of the T1-V set, T2A-V set and T2B-V set plotted in Figures 6 to 8, respectively. Hence, the T2B-V set has similar residual strengths within the post-peak range and thus similar ductility as the T2A-V set.

For the T2C group of CFST specimens filled with Type 2C concrete, the axial performance was significantly lower than the other three groups. This can be seen from the significantly lower average values of P_peak, P_trough and P_0.06 of the T2C-S set and the T2C-V set compared to the corresponding values of the other groups, as depicted in Figures 10 to 12. With or without compaction applied, the P_peak values of the CFST specimens in this group are the lowest among all the four groups of CFST specimens tested. This was due to the rather low compaction index of 0.904 and 0.953 under the uncompacted and compacted conditions, respectively, of the Type 2C concrete. In fact, as depicted in Table 3, the uncompacted strength of the Type 2C concrete was only 54.7 MPa, which was equal to 67.5% of that of the Type 1 concrete. However, the reductions of the average values of P_trough and P_0.06 due to the use of the Type 2C concrete were relatively small. As can be seen from the axial load–strain curves plotted in Figure 9, the post-peak residual strengths compared to the corresponding axial load at peak were not low at all, indicating that although the axial strength was on the low side, the ductility was actually quite good.

Discussions

Overall, the above test results of the CFST specimens reveal that the mix proportions and compaction index of the concrete infill could have significant effects on the axial performance of the CFST made with the SCC infill.

Regarding the effects of the mix proportions, it is evident that the paste volume and aggregate proportions would both affect the compaction index of the SCC infill. However, it should be noted that the compaction index of the SCC infill is not directly related to the flowability, as demonstrated by the test finding that the order of compaction index is not the same as the order of flowability. Nevertheless, the compaction index has great influence on the uncompacted strength, as indicated by the descending order of uncompacted strength of Type 1 > Type 2A > Type 2B > Type 2C, which matches exactly with the descending order of compaction index. Among the four concrete mixes, the Type 1 mix has the larger paste volume of 40.8%, whereas the other concrete mixes have the smaller paste volume of 36.7%. Quite possibly, the Type 1 mix has the highest compaction index on average mainly because of its larger paste volume. The aggregate proportions also have certain effects. Among the Type 2 mixes, the Type 2A mix has on average a larger aggregate size than the Type 2B and Type 2C mixes and thus a smaller aggregate surface area. In theory, with the same paste volume, a smaller aggregate surface area would lead to a larger paste film thickness (Kwan and Li, 2012, 2014). So, quite possibly, the Type 2A mix has a higher compaction index on average than the Type 2B and Type 2C mixes mainly because of its larger paste film thickness. It is postulated herein that both the paste volume and paste film thickness have significant effects on the compaction index.

Regarding the effects of the compaction applied to the concrete infill, it is observed that if both the compaction index under the uncompacted condition and the compaction index under the compacted condition are high and close to each other, as for the Type 1 mix tested herein, then the compaction applied should have little effects on the strength of the concrete and the axial performance of the CFST made with the concrete. However, it should be noted that both types of compaction indices should be considered. Just saying that the concrete is well compacted when the uncompacted density is close to the compacted density, as in the European Standard BS EN 12350-4 (2009), could be quite misleading. If the compaction index under the compacted condition is significantly lower than 1.0, as for the Type 2C mix tested herein, then even with good compaction applied, the strength of the concrete and the axial performance of the CFST made with the concrete would still be adversely affected. It is therefore suggested that both the compaction index under the uncompacted condition and the compaction index under the compacted condition should be determined to evaluate the suitability of the concrete mix for being placed as SCC with no compaction applied.

Finally, it is noted that the paste volume and aggregate proportions of the concrete infill also have significant effects on the post-peak behaviour, especially the post-peak residual strength and ductility, of the CFST made with the concrete. Among the four groups of CFST, the T1 group has the lowest P_trough to P_peak ratio, indicating that this group has proportionally the lowest residual strength at the post-peak stage and thus the lowest ductility. This is a new discovery and the exact reasons for such effects are yet to be investigated. As the major differences between the Type 1 mix and the other concrete mixes are the larger paste volume and larger paste film thickness of Type 1 mix, it is postulated herein that such effects are related to the aggregate–aggregate and aggregate–paste interactions within the concrete mix. At the post-peak stage, micro-cracks in the paste and aggregate–paste interface debonding should have occurred, and the major mechanism left to provide residual strength is the aggregate interlocking action under confinement (Piotrowska et al., 2014). The larger paste volume and paste film thickness in the Type 1 mix have caused the average inter-particle distance between the aggregate particles to be larger than those in the other concrete mixes. This might have significantly reduced the aggregate interlocking action at the post-peak stage and thus adversely affect the post-peak residual capacity.

Conclusion

This article has presented a pilot study to investigate the influences of the SCC mix proportions on the axial performance of CFST through experiments, with the findings summarized as follows:

A ‘compaction index’ being equal to the apparent wet density to theoretical wet density ratio, that is, $(M_{c} / V_{c}) / ρ_{con}$ , is introduced to evaluate the compaction level of the concrete infill for CFST columns.

The compaction index of the SCC is not directly related to the flowability, but a higher compaction index will lead to a better axial performance for the CFST column made with SCC under uncompacted condition.

In a certain range, a higher paste volume will normally bring a higher compaction index for the SCC. When the paste volume is constant, the SCC with a smaller aggregate surface area would have a larger paste film thickness, as well as a higher compaction index.

In real construction, determining both the compaction index under the uncompacted condition and the compaction index under the compacted condition would be a good engineering practice to evaluate the suitability of the concrete mix for being placed as SCC without any extra means of compaction.

When the compaction index is equal to or very close to 1.00, a larger paste film thickness in the SCC mix means a larger average inter-particle distance between the aggregate particles, potentially reducing the aggregate interlocking action at the post-peak stage and thus adversely affecting the post-peak residual capacity. Therefore, in real construction, if the application of vibratory compaction is feasible for the CFST structures, choosing an SCC mix with relatively small paste film thickness would improve the ductility of the CFST structures. However, the paste film thickness cannot be overly reduced, as exemplified by the low axial strengths of the CFST specimens from the T2C group.

Inspired by this study, a systematic correlation between the paste film thickness and the lateral dilation behaviours triggered by the aggregate–aggregate and aggregate–paste interactions of concrete under triaxial compressive state is worth being further investigated in the future studies.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: The authors gratefully acknowledge the financial support provided by National Natural Science Foundation of China (Project Nos. 51808134 and 51608131), the Guangdong Science and Technology Department (No. 2017A010103030), the Guangzhou Science and Technology Department (No. 201904010164), Featured and Innovative Project for Colleges and Universities of Guangdong Province (Project No. 2017KTSCX061), Pearl River S&T Rising Star Program of Guangzhou City (Project No. 201906010064) and ‘One-Hundred Young Talents Plan of Guangdong University of Technology’ (Project Nos 220413226 and 220413632 and 220413508).

ORCID iD

JJ Zeng

References

BS 18-87 (1987) Method for tensile testing of metals.

BS EN 12350-11 (2010) Testing fresh concrete: part 11: self-compacting concrete – sieve segregation test.

BS EN 12350-2 (2009) Testing fresh concrete: part 2: slump-test.

BS EN 12350-4 (2009) Testing fresh concrete: part 4: degree of compactability.

BS EN 12350-8 (2010) Testing fresh concrete: part 8: self-compacting concrete – slump-flow test.

Chen

(2008) Examples of Concrete-Filled Steel Tube Arch Bridges (II). Beijing, China: People Traffic Publishing House (in Chinese).

Fam

Rizkalla

(2001) Confinement model for axially loaded concrete confined by circular fiber-reinforced polymer tubes. ACI Structural Journal 98(4): 451–461.

GB 175-2007 (2007) Common Portland cement (in Chinese).

GB/T 27690-2011 (2011) Silica fume for cement mortar and concrete (in Chinese).

10.

Han

(2010) Seismic performance of CFST column to steel beam joint with RC slab: experiments. Journal of Constructional Steel Research 66(11): 1374–1386.

11.

Han

Yao

(2004) Experimental behaviour of thin-walled hollow structural steel (HSS) columns filled with self-consolidating concrete (SCC). Thin-Walled Structures 42(9): 1357–1377.

12.

Han

Yao

, et al. (2006) Further study on the flexural behaviour of concrete-filled steel tubes. Journal of Constructional Steel Research 62(6): 554–565.

13.

Han

Ren

(2010) Tests on inclined, tapered and STS concrete-filled steel tubular (CFST) stub columns. Journal of Constructional Steel Research 66(10): 1186–1195.

14.

Han

Yao

Zhao

(2005a) Tests and calculations for hollow structural steel (HSS) stub columns filled with self-consolidating concrete (SCC). Journal of Constructional Steel Research 61(9): 1241–1269.

15.

Han

You

Lin

(2005b) Experimental behaviour of self-consolidating concrete (SCC) filled hollow structural steel (HSS) columns subjected to cyclic loadings. Advances in Structural Engineering 8(5): 497–512.

16.

Han

Zhao

(2014) Concrete-Filled Tubular Members and Connections. London: CRC Press.

17.

Kwan

AKH

(2012) Combined effects of water film thickness and paste film thickness on rheology of mortar. Materials and Structures 45(9): 1359–1374.

18.

Kwan

AKH

(2014) Combined effects of water film, paste film and mortar film thicknesses on fresh properties of concrete. Construction and Building Materials 50: 598–608.

19.

Kwan

AKH

(2015) Effects of superplasticizer type on packing density, water film thickness and flowability of cementitious paste. Construction and Building Materials 86: 113–119.

20.

Han

(2011) Seismic performance of CFST column to steel beam joints with RC slab: analysis. Journal of Constructional Steel Research 67(1): 127–139.

21.

Han

(2012) Seismic performance of CFST column to steel beam joint with RC slab: joint model. Journal of Constructional Steel Research 73: 66–79.

22.

Ling

Kwan

AKH

(2018) Filler technology for low-carbon high-dimensional stability concrete. Construction and Building Materials 163: 87–96.

23.

Liu

Wang

Chan

(2017) Behaviour of concrete-filled cold-formed elliptical hollow sections with varying aspect ratios. Thin-Walled Structures 110: 47–61.

24.

Han

Zhao

(2010a) Fire performance of self-consolidating concrete filled double skin steel tubular columns: experiments. Fire Safety Journal 45(2): 106–115.

25.

Zhao

Han

(2009) Fire behaviour of high strength self-consolidating concrete filled steel tubular stub columns. Journal of Constructional Steel Research 65(10–11): 1995–2010.

26.

Zhao

Han

(2010b) Testing of self-consolidating concrete-filled double skin tubular stub columns exposed to fire. Journal of Constructional Steel Research 66(8–9): 1069–1080.

27.

Martínez-Lage

Martínez-Abella

Vázquez-Herrero

, et al. (2012) Properties of plain concrete made with mixed recycled coarse aggregate. Construction and Building Materials 37: 171–176.

28.

Maybury

JCM

Binhowimal

(2017) Fillers to lessen shear thickening of cement powder paste. Construction and Building Materials 142: 268–279.

29.

Okamura

Ouchi

(2003) Self-compacting concrete. Journal of Advanced Concrete Technology 1(1): 5–15.

30.

Ouyang

Kwan

AKH

(2018) Finite element analysis of square concrete-filled steel tube (CFST) columns under axial compressive load. Engineering Structures 156: 443–459.

31.

Ouyang

Kwan

AKH

, et al. (2017) Finite element analysis of concrete-filled steel tube (CFST) columns with circular sections under eccentric load. Engineering Structures 148: 387–398.

32.

Piotrowska

Malecot

(2014) Experimental investigation of the effect of coarse aggregate shape and composition on concrete triaxial behavior. Mechanics of Materials 79: 45–57.

33.

Ren

Han

Lam

, et al. (2014) Experiments on special-shaped CFST stub columns under axial compression. Journal of Constructional Steel Research 98: 123–133.

34.

Wang

Xie

(2015) Design and test of high-performance arch rib concrete of Bosideng Yangtze River Bridge. Journal of Guangxi University 40(5): 1075–1079 (in Chinese).

35.

Wong

Kwan

AKH

(2008) Packing density of cementitious materials: part 1 – measurement using a wet packing method. Materials and Structures 41(4): 689–701.

36.

Xiong

Liew

JYR

(2017a) Axial performance of short concrete filled steel tubes with high- and ultra-high- strength materials. Engineering Structures 136: 494–510.

37.

Xiong

Liew

JYR

(2017b) Behaviour of steel tubular members infilled with ultra high strength concrete. Journal of Constructional Steel Research 138: 168–183.

38.

Chen

, et al. (2018) Cyclic behaviour of double-tube buckling-restrained braces for boiler steel plant structures. Journal of Constructional Steel Research 150: 556–569.

39.

Kim

Han

, et al. (2012) Blast resistant characteristics of ultra-high strength concrete and reactive powder concrete. Construction and Building Materials 28(1): 694–707.

40.

Tao

(2008) Experimental behaviour of high performance concrete-filled steel tubular columns. Thin-Walled Structures 46(4): 362–370.

41.

Ding

Cai

(2007) Experimental behavior of circular concrete-filled steel tube stub columns. Journal of Constructional Steel Research 63(2): 165–174.

42.

Zhao

Packer

(2009) Tests and design of concrete-filled elliptical hollow section stub columns. Thin-Walled Structures 47(6–7): 617–628.

43.

Zhu

Chan

Young

(2019) Cross-sectional capacity of octagonal tubular steel stub columns under uniaxial compression. Engineering Structures 184: 480–494.

44.

Zhu

Liu

Wang

, et al. (2010) Experimental research on square steel tubular columns filled with steel-reinforced self-consolidating high-strength concrete under axial load. Engineering Structures 32(8): 2278–2286.