Compressive and tensile behaviour of alkali‐activated slag‐based concrete incorporating single hooked‐end steel fibres

The effect of single hooked‐end steel fibres, namely Dramix® 3D, on the mechanical performance of alkali‐activated slag‐based concrete (AASC) and Portland cement concrete (PCC) has been investigated. Compressive strength, modulus of elasticity, stress‐strain response under uniaxial compression and tension, splitting tensile strength and flexural strength have been evaluated. The experimental results show that AASC incorporating 3D fibres in a volume fraction of 0.75% exhibits an enhanced behaviour, under both compression and tension, in comparison to PCC incorporating the same fibre type and dosage. Although the reference mixes show similar compressive strength, 3D fibres enhance the modulus of elasticity, splitting tensile strength and flexural strength of AASC of 8.6%, 61.7% and 12.8%, respectively, while for PCC 1.1%, 42.2% and 16.1%, respectively. Three‐point bending tests show the effect of 3D fibres on the response of AASC and PCC under flexural loading. Although fibres have a limited effect on the strength corresponding to the limit of proportionality (LOP), they enhance the post‐peak behaviour, increasing the residual flexural strength and the material ductility. Finite element analysis has been performed to predict the flexural behaviour of steel fibre‐reinforced AASC (FRAASC) under flexural loading. The Concrete Damage Plasticity (CDP) model implemented in ABAQUS software can predict the flexural response of FRAASC quite accurately, although additional experimental data are needed to calibrate the model for different alkali‐activated matrix types.


Introduction
The environmental impact of the Portland cement-based concrete (PCC) production process has recently become a serious concern, requiring demanding solutions in the construction sector.Alkali-activated concrete (AASC) represents an environmental friendly alternative to PCC, as it can achieve higher mechanical and durability performance while reducing the CO2 emissions linked to its manufacturing [1].However, like PCC, AASC exhibits brittle behaviour and low tensile and flexural strengths [2].The incorporation of randomly distributed steel fibres enhances the composite tensile behaviour and ductility, thereby mitigating crack formation and propagation [3].Although steel fibres, in particular hooked-end steel fibres, are highly used in PCC composites [4], studies investigating their effect on the compressive and tensile behaviour of AASC are still limited.Bernal et al. [5,6] evaluated the mechanical and durability performance of AASC incorporating hooked-end steel fibres (length   = 25 mm and diameter   = 1 mm) in different volume fractions (0%, 0.5% and 1.5%) and compared it to PCC mixes reinforced with the same fibre type and content.It was observed that for AASC, fibre addition results in a reduction of compressive strength with the increase of the fibre dosage while enhancing its splitting tensile strength, flexural strength, and water absorption.Furthermore, the performance enhancement due to hooked-end steel fibre incorporation was more evident for AASC than PCC, demonstrating the suitability of AASC as a possible alternative to PCC.Kim et al. [7] evaluated the effect of hooked-end steel fibres (  = 30 mm,   = 0.5 mm) in different volume fractions (0-2 %) on the compressive and tensile behaviour of AASC.Compressive strength, stress-strain response under uniaxial compression and flexural performance have been investigated in their study.The compressive strength and the strain corresponding to the peak compressive stress increased with the increase of the fibre volume fraction, resulting in a more ductile failure in comparison to the unreinforced concrete.They further observed that under flexural loading, fibres have a limited effect on the flexural strength, while they improve considerably the post-peak ultimate strength with the increase of the fibre volume fraction.
Although few researchers evaluated the effect of hookedend steel fibres on the mechanical performance of AASC, studies characterising the constitutive behaviour of steel fibre-reinforced AASC (FRAASC) under compression and tension, are still missing.Evaluating the compressive and tensile response of hooked-end steel FRAASC is fundamental to deriving analytical equations for design purposes and promoting the use of this alternative construction material in practice.When the material behaviour cannot be determined experimentally, finite element (FE) analysis can be performed to model and predict the material post-cracking non-linear response.The Concrete Damage Plasticity (CDP) model implemented in ABAQUS software can be used for steel fibre-reinforced concrete, when the input parameters, i.e. the stress-strain response under compression and tension, are known, whether obtained experimentally or analytically.
This study evaluates the compressive and tensile behaviour of AASC and PCC mixes incorporating hooked-end steel fibres at 0.75% volume fraction.Compressive strength, modulus of elasticity, stress-strain behaviour under uniaxial compression and uniaxial tension, splitting tensile strength and flexural strength have been investigated.The stress-strain responses obtained experimentally have been used as input parameters for the CDP model available in ABAQUS software to predict the behaviour of steel FRAASC under flexural loading.The goal is to evaluate the suitability of hooked-end steel fibres for alternative concrete matrices and understand their effect on the mechanical performance of alkali-activated slag-based concrete.

2
Experimental programme

Materials and mix proportions
Portland cement (CEM I 42.5R) and ground granulated blast furnace slag (GGBS) supplied by HeidelbergMaterials and Ecocem Benelux B.V., respectively, have been used in this study.The chemical composition of each binder is shown in Table 1.Single hooked-end steel fibres, namely Dramix ® 3D 65/60 BG, with a length of 60 mm and an aspect ratio of 65 (i.e. the ratio between the fibre length (  ) and the fibre diameter (  )) have been added to each concrete matrix in a volume fraction of 0% (reference) and 0.75%.

Casting and curing conditions
For both concrete matrices, the same mixing procedure has been followed.First, the dry components (binder and fine and coarse aggregates) were mixed in a 250-litre mixer for 90 s.Then the liquid components, i.e. water for PCC and the alkaline solution for AASC were added and the mixing was prolonged for an additional 2 min.Finally, steel fibres were added and the concrete was further mixed for a total of 7 minutes.For each mix, 15 cylinders with a diameter of 150 mm and a height of 300 mm and 6 beams with dimensions of 150 mm x 150 mm x 550 mm were cast.Before demoulding, the samples were cured for 24 hours in the moulds at room temperature and covered to avoid moisture loss.After demoulding, PCC samples have been cured underwater ((20±2) °C), while AASC samples have been wrapped in foil and stored in a climate room at (20±2) °C and 65% relative humidity until the testing date.After 27 days of curing, a notch of 5 mm width and 25 mm depth has been sawed at the mid-span of the side face of each beam according to EN 14651.

Testing procedure
For each mix, both the compressive and the indirect tensile behaviour have been investigated.Table 3 summarised the performed tests, the relative EN standard and the number and type of samples used.The stress-strain response under uniaxial compression was evaluated using an MTS compression-tension testing machine with 2500 kN capacity.A constant loading head displacement rate of 0.30 mm/min has been applied and the axial plate-to-plate deformation was evaluated with four external linear variable differential transformers (LVDTs).The test setup chosen for the determination of the stress-strain response under compression does not allow the evaluation of the modulus of elasticity (no deformation in the central section of the cylinder can be evaluated) but provides a good reading of the post-peak descending branch [10], where the effect of fibres is more pronounced.In addition, for AASC-REF and AASC-3D75, direct tensile tests have been performed after 7 days on 3 notched prisms of dimensions 100 mm x 100 mm x 260 mm, with a net cross-section of 60 mm x 100 mm.Each sample was glued to the machine steel plates before a constant tensile with a rate of 0.03mm/min was applied.Six different LVDTs -one at each corner and two on the notched surfaces -were placed in the central section of the sample (total measuring length = 50 mm) to record the load-deformation curve.

Finite element simulation
The flexural behaviour of steel FRAASC under three point bending has been simulated using finite element simulation with ABAQUS software.In this study, the Concrete Damage Plasticity Model (CDPM) implemented in ABAQUS has been used to simulate the behaviour of steel FRAASC under flexural loading (3PBT).The composite is considered an isotropic continuum material, i.e. the fibres are not directly included in the model, but their effect is considered in the material properties.The key CDPM parameters are given in Table 4. Input parameters such as density, compressive strength, modulus of elasticity and stress-strain response under compression and tension of steel FRAASC were obtained experimentally.

Compressive strength and modulus of elasticity
Figure 1 shows the 28-day mean values of compressive strength and modulus of elasticity of AASC and PCC with and without hooked-end steel fibres.
The effect of fibre incorporation on these mechanical properties differs depending on the matrix type.For AASC, hooked-end steel fibres enhance both the compressive strength and the modulus of elasticity of the plain matrix, resulting in an increase of 18.0% and 8.6%, respectively.For PCC, hooked-end steel fibres have a negative effect on the compressive strength (-11.3% in comparison to the reference) and a negligible effect on the modulus of elasticity (+1.1% in comparison to the plain matrix).
The enhanced behaviour of AASC incorporating steel fibres in comparison to PCC can be linked to the higher shrinkage characterising alkali-activated slag-based concretes, which improves the fibre-matrix bond [11] and consequently the material performance.

Compressive stress-strain response under uniaxial loading
Figure 2 shows the 28-day mean experimental and normalised stress-strain responses under uniaxial compression for AASC and PCC with and without steel fibres.
Although PCC-REF shows a higher peak stress than AASC-REF, 41.22 MPa and 38.32 MPa, respectively, the incorporation of steel fibres has a different effect on the peak stress depending on the matrix type.
For AASC, the incorporation of steel fibres at 0.75% volume fraction enhances the peak stress in comparison to the reference of 17.1%, while for PCC it has a limited effect (-1.4% in comparison to PCC-REF).The same can be seen for the strain corresponding to the peak stress The incorporation of steel fibres enhances the  , of both AASC and PCC, with AASC-3D75 showing the highest increase (+22.5% in comparison to AASC-REF).A higher value of strain corresponding to peak stress, coupled with a softer post-peak descending branch of the mixes containing fibres in comparison to the references, represents a higher ductility.
The increment in compressive strength given by incorporation of hooked-end steel fibres is higher in AASC than PCC.This is because the enhancement of the fibre-matrix bond results in better crack bridging and mechanical performance.This can be seen in the normalised stress-strain curves of both AASC and PCC with and without fibres (Fig. 2b).Steel fibres incorporation has a limited effect on the pre-peak ascending branch, which is mainly dependent on the matrix compressive strength.
However, steel fibres incorporation enhances the postpeak descending branch, which shows a softer slope compared to the reference mixes.AASC-3D75 exhibits higher ductility in comparison to PCC-3D75, recognisible by the higher values of normalised stress corresponding to the same strain in the post-peak regime.Like the compressive strength and the modulus of elasticity, the higher value of splitting tensile strength of AASC-3D75 in comparison to PCC-3D75 can be linked to the higher shrinkage of the alkali-activated matrix in comparison to traditional cement-based concrete.The compressive stresses around the fibres developed by the matrix undergoing shrinkage enhances the fibre-matrix bond [11] and allows for an early activation of the fibres bridging the shrinkage-induced micro-cracking.

Flexural strength at the limit of proportionality (𝐟 𝐋𝐎𝐏 ) and residual flexural strength
The fib Model Code 2010 provides a classification of steel fibre-reinforced PCC based on the flexural post-peak behaviour obtained by performing three-point bending tests (3PBTs) on notched beams according to EN 14651.From the Load-CMOD (crack mouth opening displacement) recorded experimentally, the values of the flexural strength corresponding to the Limit of Proportionality (  ) and the residual flexural tensile strength   can be derived according to the equations: where   is the load corresponding to the LOP (i.e. the maximum load for CMOD ≤ 0.05 mm),   (with j = 1, 2, 3, 4) is the load corresponding to CMOD of 0.5, 1.5, 2.5 and 3.5 mm, respectively,  and  are the span and the width of the beam, respectively, and ℎ  is the distance between the tip of the notch and the top of the beam.The residual strengths corresponding to CMOD of 0.5 mm ( 1 ) and 2.5 mm ( 3 ) have been chosen for the material classification, as they describe the material performance at the serviceability and ultimate limit states, respectively.
Figure 4 shows the stress-CMOD curves (average values) obtained for AASC and PCC with and without hooked-end steel fibres.Hooked-end steel fibres have a limited effect on the flexural strength   , which is generally strongly dependent on the concrete strength [3].When the entire stress-CMOD curves are considered, it can be easily seen that AASC and PCC incorporating hooked-end steel fibres has a similar behaviour.
However, AASC with and without fibres show higher values of   and residual flexural strengths  1 and  3 corresponding to CMOD values of 0.5 mm and 2.5 mm, respectively, as shown in Fig. 5.
The behaviour shown in Fig. 4 and Fig. 5 can be seen also in the classification proposed by the fib Model Code 2010 and shown in Table 5.According to the fib Model Code 2010, steel fibre-reinforced concrete can be classified with a number and a letter.The number represents the characteristic residual flexural strength  1 , while the letter corresponds to specific values of the  3 / 1 .The characteristic values  1 and  3 are obtained following the equation: where   (with j = 1 and 3) represent the average residual flexural strength corresponding to CMOD of 0.5 and 2.5 mm, respectively, and can be derived according to Eq. ( 2) and  represents its standard deviation.

Direct tension tests
Figure 6 shows the load-displacement curves and the peak load for AASC-REF and AASC-3D75.
Hooked-end steel fibres have a limited effect on the peak tensile stress.However, fibres enhance the post-peak behaviour, increasing the material ductility.Nevertheless, direct tension tests performed on steel fibre-reinforced composites can only give an overall indication of the material response, as the value of the peak stress is highly influenced by the amount and orientation of the fibres in the cross-section.A higher amount of fibres aligned in the direction of the tensile load will result in higher values of tensile strength.

Finite element simulations
Figure 7 shows the tensile damage parameter and the maximum principal stresses in the finite element simulation for AASC-3D75.
Under three-point bending, the tensile stresses concentrate at the top of the notch and as the load increases, a crack starts generating and propagating.The crack propagation is directly correlated to the tensile damage parameter represented in Fig. 7a.Once the peak load is achieved, the beam does not fail abruptly, as the presence of the hooked-end steel fibres slows the crack propagation and allow redistribution of stresses in the beam (Fig. 7b).
Figure 8 shows the experimental and predicted load-displacement curves for steel fibre-reinforced alkali-activated concrete.Although the CDPM is implemented in ABAQUS for plain cement-based concrete, it can also be used to predict the flexural behaviour of steel fibre-reinforced alkali-activated concrete.Additional experimental data, in particular stress-strain curves under uniaxial tension, and further parameter calibration are needed to better predict the behaviour of steel fibre-reinforced AASC.

Conclusions
This study evaluated the compressive and tensile behaviour of AASC and PCC incorporating hooked-end steel fibres in a volume fraction of 0.75%.Compressive strength, modulus of elasticity, stress-strain under uniaxial compression and tension, splitting tensile strength and flexural strength have been evaluated and the following conclusions can be drawn: • Hooked-end steel fibres enhance the compressive strength and the modulus of elasticity of AASC by 18% and 8.6%, respectively, in comparison to the reference concrete.In PCC mixes, the addition of hooked-end steel fibres results in a decrease in compressive strength (-11.3%) and a slight increase in the modulus of elasticity (+1.1%) in comparison to the concrete without fibres.

•
Hooked-end steel fibres have generally a limited effect on the pre-peak ascending branch of the stress-strain curve under uniaxial compression, while they enhance the peak stress, the corresponding strain and the post-peak residual stress at higher strain values.If the enhancement of peak stress and corresponding strain is significant for AASC, +17.1% and +22.5%, respectively, for PCC it is quite limited (-1.4% and +5.4%, respectively).This, combined with higher stress at higher post-peak strain rates, demonstrates the better behaviour of FRAASC than FRPCC when subjected to uniaxial compressive loading.

•
For both AASC and PCC, the addition of hookedend steel fibres enhances the splitting tensile strength.However, the increase of strength in comparison to the plain matrix is higher for AASC than PCC (+61.7% and +42.2%, respectively).

•
The flexural strength corresponding to the limit of proportionality   increases with the incorporation of a 0.75% volume fraction of hooked-end steel fibres.Despite showing similar flexural behaviour, fibre-reinforced AASC shows higher values of   and  3 in comparison to PCC-3D75.

•
When subjected to uniaxial tension loading, AASC exhibits a brittle behaviour, which is represented by a softening post-peak descending branch.Hooked-end steel fibres slightly affect the peak stress, while enhancing the post-peak behaviour and load-bearing capacity of the material.

•
AASC incorporating hooked-end steel fibres represent a valuable alternative to PCC, as it exhibits improved mechanical performance in comparison to PCC.This can be related to the alkali-activated slag-based matrix's higher shrinkage in comparison to PCC, which leads to a better fibre-matrix bond and improved mechanical behaviour.

•
The Concrete Damage Plasticity Model (CDPM) implemented in the ABAQUS software can be used  to predict the flexural behaviour of steel fibre-reinforced AASC.Additional experimental results, in particular compressive and tensile stress-strain response under uniaxial loading, are needed to better calibrate the model and further apply it to any steel fibre-reinforced alkali-activated slagbased concrete.
This study demonstrates that alkali-activated slag-based concrete incorporating hooked-end steel fibres can outperform FRPCC from a mechanical point of view.However, additional studies characterising the behaviour of steel fibre-reinforced AASC, focusing in particular on the stressstrain behaviour under uniaxial compression and tension, are needed to derive analytical and numerical models for design and field applications.

Figure 1
Figure 1 28-day mean compressive strength (′  ) and modulus of elasticity (  ) of AASC and PCC with and without steel fibres

Figure 3
Figure 3 shows the 28-day mean splitting tensile strength  , of both AASC and PCC with and without steel fibres.Although PCC-REF has a slightly higher value of splitting tensile strength than AASC-REF (3.27 MPa and 3.16 MPa, respectively) AASC-3D75 exhibits a higher splitting tensile strength than PCC-3D75 (5.11 MPa and 4.65 MPa, respectively).Steel fibres provide an enhancement of splitting

Figure 3 Figure 2
Figure 3 28-day mean splitting tensile strength ( , ) of AASC and PCC with and without steel fibres

Figure 5 Figure 6 7 -Figure 4
Figure 5 28-day mean flexural strength corresponding to the limit of proportionality (  ) and the residual strengths  1 and  3 for AASC and PCC with and without steel fibres

Figure 7
Figure 7 Tensile damage parameter (a) and maximum principal stress distribution (b) corresponding to the limit of proportionality for AASC-3D75

Figure 8
Figure 8 Experimental and predicted load-displacement curves for AASC-3D75

Table 2
Mix proportions of PCC and AASC

Table 3
Tests performed to evaluate the compressive and tensile behaviour of PCC and AASC with and without steel fibres

Table 4
General CDPM parameters

Table 5
Classification of the different concrete mixes according to the fib Model Code 2010