Long term performance of recycled concrete beams with different water–cement ratio and recycled aggregate replacement rate

Abstract

The purpose of this study is to reveal the service performance of recycled aggregate concrete (RAC) components for different values of water−cement ratio and replacement rate of recycled coarse aggregate (RCA). Generally, the concrete strength decreases with the increase of the replacement rate of RCA, in order to meet the strength requirements when changing the replacement rate of RCA, it is necessary to change the water−cement ratio at the same time. Therefore, the axial compressive strengths of prism with 25 mix proportions, the short-term mechanical properties and long-term deformation properties of reinforced concrete beams were tested respectively by changing water−cement ratio and RCA replacement rate. The bearing capacity and the strain nephogram of samples under different loads were obtained using the Digital Image Correlation (DIC) method, and a self-made gravity loading experimental device was used for long-term deformation investigation. Results showed that the damage pattern of RAC was the same as that of natural aggregate concrete (NAC), but the brittleness was more pronounced. The brittleness of concrete before failure can be reduced more effectively by adjusting the replacement rate of RCA than by adjusting the water−cement ratio. The water−cement ratio has an evident influence on the axial compressive strength and early creep of concrete, while the replacement rate of RCA has a remarkable effect on the long-term deformation of the concrete beams.

Graphical abstract

Keywords

recycled concrete / beam / the replacement rate of recycled coarse aggregate / water–cement ratio / digital image correlation

1 Introduction

The application of recycled coarse aggregate (RCA) obtained by crushing waste concrete in new fresh concrete [1] is necessary and beneficial. Applying RCA in concrete structures has been proved feasible [2]. This is an effective solution to construction waste. However, compared with natural coarse aggregates (NCA), RCA has inherent defects due to many micro-cracks occurring in the RCA crushing process. This results in RCA usually having greater porosity and absorption, lower apparent density and strength than NCA [3]. The shape of RCA is often more angular and sharper, and there is also an interfacial transition zone (ITZ) of the RCA between new and old mortar on the surface [4]. These differences lead to many differences in the performance between RCA and NCA [5,6] and affect the practical application of recycled aggregate concrete (RAC). In addition, the old mortar attached to the surface of RCA often has more microcracks [7]. These microcracks lead to differences between natural aggregate concrete (NAC) and RAC in strength, shrinkage and creep [8–10].

The work of some researchers has made RAC widely used in building structures [11–13]. Beams are essential components in building structure, so it is significant to study the service performance of RAC beams for practical engineering applications. The replacement rate of RCA has a significant influence on the compressive strength of RAC at each curing age [14]. Furthermore, the compressive strength of RAC with different replacement rates of RCA can meet specific design requirements by adjusting the water−cement ratio. Chen et al. [15] found that the prism compressive strength of RAC at 28-d curing age increased slightly with the increase of replacement rate of RCA. In addition, RAC can achieve more satisfactory results in mechanical properties through reasonable and technical methods than NAC beams [16]. Etxeberria et al. [17] found that the performance of RAC beam is not significantly reduced compared with that of NAC beam when the replacement rate of RCA is less than 25%. Moreover, they also found that after further durability measures are taken, RCA can be used in concrete structures. Adam and Knaack [18] studied the flexural and shear properties of RAC beams and found that RCA has little effect on flexural and shear strength of beams. Choi and Yun [19] produced beams with the RCA replacement rate of 0%, 50%, and 100%, respectively. They found that the initial flexural stiffness and maximum flexural strength of beams with different aggregate types are similar, while the maximum flexural strength of NAC beams is 20% higher than that of RAC beams. Arezoumandi et al. [20] produced eight beams with different reinforcement ratios and concrete types and tested them all under four-point supported loads. They found that the RAC beams have considerable flexural strength, but their deflection is 13% greater than that of the NAC beams. Yang et al. [21] conducted a regular section bending performance test on the RAC beams. The bending performance test found that the deflection of RAC beams decreases with the increase of longitudinal steel reinforcement ratio and concrete strength. The measured deflection of the RAC beam is about 10% larger than the calculated value of GB50010-2010 [22]. According to the above research results, it can be found that the bearing capacity and deflection of concrete beam is related to many factors, such as the reinforcement ratio, the replacement rate of RCA.

The crack width in RAC beams is more significant than that in NAC beams, but the crack spacing is smaller than that of the NAC beams [23], which often results in errors when visual observations are performed. Some researchers have used digital image correlation (DIC) to analyze the bending failure process of concrete beams [24–26]. Huang et al. [27] used DIC to test concrete deformation and crack propagation under uniaxial compression. Li et al. [28] used DIC to study the crack propagation process of RAC under uniaxial compression and compared the generation and propagation of microcracks in RAC and NAC, revealing different behaviors. Various researchers have shown that DIC is highly accurate [29]. Therefore, DIC can be used to analyze the strain field of the concrete surface and so to study its cracking process while under load.

Up to now, RAC performance has been investigated by changing the water−cement ratio or the replacement rate of RCA [30]. Generally, the concrete strength decreases with the increase of the replacement rate of RCA. In practical engineering, the design of structure is often based on the concrete strength. Therefore, in order to meet the strength requirements when changing the replacement rate of RCA, it is necessary to change the water−cement ratio at the same time. The gravity loading method based on the lever principle has many advantages, such as stable loading, little rusting of springs and nuts, and lower stress relaxation [31,32]. Therefore, this experiment followed the lever principle and used gravity to load concrete beams with similar strength levels, different RCA replacement rates and water−cement ratios. In addition, a self-made loading device and DIC method were applied to study the bearing capacity and mechanical properties of RAC beams with different RCA replacement rates and water−cement ratio under the long-term continuous load in an intuitive way. The mechanical properties and time-varying properties of concrete beams with different mix proportions were obtained through this experiment. In addition, the fitting formula of the relative deflection of RAC beams with mix proportion and time was established based on the experimental data to provide a reference for the long-term deformation prediction of similar RAC beams.

2 Experimental investigation

2.1 Material preparation

The cement used in this experiment was P42.5 Conch ordinary Portland cement (OPC). The specific surface area of cement was 350 m²/kg, which meets the requirements of Chinese code GB 175-2007. The physical properties and chemical composition of this cement are shown in Tab.1 and Tab.2, respectively. The fine aggregate was machine-made sand, and its properties are shown in Tab.3. The NCA was 5–25 mm continuous graded limestone gravel. RCA was the crushed concrete sourced from the grounds of Guangxi University. The coarse aggregate was washed repeatedly with water to reduce the mud content, and then screened. After being screened, the particle size range of the RCA was 5–20 mm. The physical properties of coarse aggregate are shown in Tab.4. The gradation curves of coarse aggregate and fine aggregate are shown in Fig.1.

The concrete mix proportion was designed according to Mix Proportion Design of Ordinary Concrete (JGJ 55-2011), and the RCA calculation formula of the water−cement ratios was based on different replacement rates of aggregate quality proposed by Guo et al. [33]. Given that the bulk density of RCA is different from that of NCA, and RCA often has higher water absorption than NCA, more water had to be added in the designed RAC, which can lead to a significant deviation between the actual water−cement ratio of RAC and the nominal water−cement ratio. Therefore, the absolute volume method proposed by Lian and Li [34] was used to calculate the amount of each material. According to the absolute volume method, the RCA replacement rates were 0%, 25%, 50%, 75%, and 100%. In addition, the ratio between the volume of cement paste and volume of aggregate (VP/VA) was 0.47, matching the maximum VP/VA ratios of different grades of concrete recommended by Lian and Li [34], the sand rate was set as 40% according to the maximum particle size of the coarse aggregate. The slump of the prepared concrete was about 170 mm. To further study the influence of the replacement rate of RCA and water−cement ratio changed simultaneously on the service performance of RAC, samples with different water−cement ratios were prepared for each replacement rate of RCA. 25 mix proportions were applied and are shown in Tab.5. The mix proportion of concrete used for casting beams is the same as NC0, RC25, RC50, RC75, and RC100.

HRB400 hot-rolled ribbed reinforcement with a diameter of 12 mm was used for longitudinal tensile reinforcement at the bottom of the beams. HRB335 hot-rolled ribbed reinforcement with a diameter of 8 mm was used for the upper reinforcement, and HPB300 plain round reinforcement with a diameter of 6 mm was used for stirrup. The mechanical properties of reinforcement are shown in Tab.6.

2.2 Experiment samples

To explain the creep characteristics of the RAC beams and NAC beams, 75 samples of 100 mm × 100 mm × 300 mm concrete prism were prepared and their axial compressive strengths were tested. In addition, 10 concrete beams with a section size of 100 mm × 150 mm and with 2200 mm length were cast to test bearing capacity and long-term deformation performance. The mechanical properties of reinforcement are shown in Tab.6, and beam sample details and loading locations are shown in Fig.2. In addition, several horizontal and vertical bars were set in the shear section of the beam to separate the reinforcement cage from the template. These bars were located between the stirrups, and the properties of these bars were the same as those of stirrups. The length of these horizontal bars was equal to the width of the cross-section of the beam and the length of these vertical bars was equal to the height of the cross-section of the beam.

2.3 Experiment content

2.3.1 Experiment of mechanical properties of concrete prism

The RMT-201 rock press of Guangxi University for loading was used for the experiment. The Industrial Camera (14 million pixels, Beijing JingHang, JHSM1400f) and 800 W halogen illuminator were used to collect DIC data. The model of the displacement meter is CW-341 and the brand is JingZhun. The accuracy of the displacement meter is 1 micron and the measurement range is 10 mm. Before the experiment, the samples were removed from the bench in the concrete curing room to ensure they were dry. The displacement gauge bracket was fixed on the sample so that the bracket was located in the middle of the sample, and the distance between the upper and lower brackets was kept at 100 mm. The surface of the samples was polished by grinding and then sprayed with a matte paint to make speckles. The sample was placed on the loading platform of the RMT-201 rock press, and was made aligned with the centerline of the platform; then the illuminator and the industrial camera were placed in the designated positions. The industrial camera was installed and debugged for assurance of capture of precise and reliable pictures, as shown in Fig.3.

At the beginning of the experiment, the load was preset to 30% of the peak load at the loading rate of 0.004 mm/s. The purpose of this step was to eliminate the gap between the sample and the RMT-201 rock press. The sample was loaded at the rate of 0.002 mm/s until it was crushed. In the loading process, the displacement meter data were collected at the same time, and the industrial camera was used for DIC photography. Real-time photos were shot during the loading process with intervals of 1 s for analysis of the instantaneous displacement and strain of the sample surface. Unlike those previous contact deformation measurement methods by using displacement gauges or strain gauges, DIC can perform non-contact measurement and measure the deformation of the entire sample’s surface.

2.3.2 Experiment of mechanical properties of concrete beams

The mechanical properties of concrete beams will affect their long-term deformation. To investigate the effects of the water−cement ratio and replacement rate of RCA on the mechanical properties of concrete beams, the bending performance of 5 concrete beams with different mix proportions were tested. In this experiment, two-point symmetrical loading was put on the sample. The loading method was to fix the steel I-beam and the loading platform using steel wire rope, then to place a sample between the steel beam and the platform, and to load the sample with two identical hydraulic jacks. The Schematic diagram of experiment device is shown in Fig.4. The jacks used a digital display hydraulic meter to read the real-time pressure values and record the real-time load values during the loading process. Before the experiment, the jack needs to be calibrated using Hua Long WAW-600 computer-controlled universal testing machine to obtain the proportional relationship between the jack pressure value and the load value. During the loading process, DIC was used to obtain the nephogram of the sample surface’s strain, and to obtain the displacement of the sample. The surface of the sample had to be polished for better quality imaging, and then the matte black spray paint was used to make speckles. Two industrial cameras (14 million pixels) were placed in front of the sample to take real-time photos. The photo interval was 1 s. The DIC measurement area was from the left support to the proper support of the sample. The DIC image analysis parameters were as follows: facet size was 50 pixels, the dot pitch was 5 pixels, and strain tensor neighborhood was 5. In addition, displacement meters were installed to obtain the deflection of the sample in the middle of the span. Displacement meters were located at both points of the sample support and also in the middle of the span.

2.3.3 Time-dependent performance experiment of concrete beam

In this experiment, 5 concrete beams with different RCA replacement rates and water−cement ratio were used for a 394-d long-term time-dependent performance experiment, and industrial cameras (14 million pixels) were used to take 394 d of continuous photos of the samples for DIC image processing. Since multiple samples were loaded at the same time, both ends of the samples were clamped with two steel pipes and tied with ropes to prevent instability. The anti-instability steel pipe used was a seamless steel pipe with a diameter of 78 mm and a thickness of 5 mm. The two ends of the steel pipe were respectively embedded in the holes of the concrete floor and the upper platform. The experiment device diagram is shown in Fig.5(a). All steel pipe surfaces were coated with anticorrosive paint. The cube concrete blocks with a side length of 150 mm were stacked at both ends of the bottom sample, and then the pine blocks with a side length of 10 cm were placed on the cube concrete blocks to support the 5 upper samples. In Chinese code GB50010-2010 [22], according to the severity of corrosion medium suffered by reinforced concrete structures, the environmental categories of concrete structures are divided into five categories. In this paper, the environment category is regarded as the third category, and the maximum crack width is 0.2 mm. Since buildings can sustain cracks during regular use, the constant load value was calculated according to the requirements of the maximum crack width and maximum deflection specified in Chinese code GB50010-2010, so that under this calculated load value, both the crack width and deflection of the sample were satisfied according to GB50010-2010. The constant load value calculated according to GB50010-2010 was 21.36 kN. Four-point bending loading was applied to the samples, and the distance between the loading points was 600 mm. The four concentrated force positions of each sample were supported by pine blocks with a side length of 10 cm. These pine blocks not only reduced the stress concentration of the samples but also provided the space for the samples’ deformation. After all of the samples were installed, the waste concrete test blocks were used to apply the load to the samples one by one. The weight of a single concrete test block was 2 kN, and a concrete test block was added every two minutes until the calculated load was reached. After loading, considering the influence of the weight of waste concrete test block and I-section steel, the load on each sample was calculated. The results are shown in Fig.5(b). The DIC image acquisition device was used to take photos of the sample every day, and then the daily displacement field and strain field of the sample were obtained through computer analysis and processing.

3 Results and discussion

After the samples were prepared, the following experiment was carried out, and the experiment results are discussed in this paper.

3.1 Experiment of mechanical properties of a concrete prism

3.1.1 Axial compressive strength

In this experiment, the axial compressive strengths of experiment samples at 28 d-curing age with replacement rates of RCA at 0%, 25%, 50%, 75%, and 100% were measured. The strength values of the samples were obtained from the experiment, and a two-dimensional nephogram was drawn and is shown in Fig.6, which can clearly explain the influence of the water−cement ratio and the replacement rate of RCA on the axial compressive strength of RAC. In Fig.6, the x-axis represents the replacement rate of RCA and the y-axis represents the water−cement ratio. The average strength value of each sample is illustrated with a circle on the graph, and the strength of the sample with the same mix proportion of the experiment beam is marked with a black circle for better analysis of the experiment result.

Fig.6 indicates that the overall axial compressive strength of RAC gradually increases with the decreasing water−cement ratio and replacement rate. This may be because reducing the water−cement ratio can significantly reduce the concrete porosity and improve the concrete strength. The surface of RCA often adhered with old mortar, which reduced the bonding strength between fresh cement paste and aggregate. Furthermore, RCA often had microcracks during crushing, and these microcracks expanded under stress. Therefore, the strength of RAC decreased with the increase of the replacement rate of RCA. Besides, it is worth noticing that the water−cement ratio has a more significant effect on the change of axial compressive strength than the replacement rate of RCA. For example, the point is selected where the water−cement ratio on the two-dimensional nephogram is 0.42, and the replacement rate of RCA is 25% for analysis. It can be seen that the fixed replacement rate of RCA is 25% when the water−cement ratio changes from 0.42 to 0.46, the strength of this point changes from 45.2 to 36.8 MPa, thus the change range is 18.6%. The fixed water−cement ratio is 0.42 when the replacement rate of RCA changes from 25% to 50%, the strength of the concrete experiment sample changes from 45.2 to 43.8 MPa, thus the change range is only 3.1%. Similar analysis can be made at other points. This particular phenomenon may be because the RCA used in this experiment was taken from road broken concrete, which had a high degree of hydration and excellent performance, so that the performance of the old ITZ between aggregate and old mortar was good. Therefore, the strength of RAC used in this experiment may mainly depend on the new ITZ between new and old mortar. The water−cement ratio has a significant influence on the performance of the new interface transition zone. As a result, the influence of the water−cement ratio on the strength is far more significant than the replacement rate of RCA.

3.1.2 Full process of uniaxial compressive stress−strain of concrete prism

DIC was used to analyze the strain change process of the samples with different RCA replacement rates during the loading process, and the NAC prism numbered NC0 was taken as an example. The strain nephogram of the NC0 sample at 28 d-curing age can be seen in Fig.7. In Fig.7, the green areas indicate that the strain value here is low, and the speckle positions on the surface of the NC0 sample have changed only slightly. The red areas indicate that the strain value here is high and the speckle positions have changed greatly, which may cause cracks to occur in the red areas. By comparing the strain nephogram of the NC0 sample with the real picture of the NC0 sample, it can be found that the DIC used in this experiment had high accuracy.

During the crushing of samples, the deformation of concrete had obvious characteristics. Therefore, this paper focuses on the analysis of these characteristics of the NC0 sample, and the characteristics of other samples, which are similar to those of the NC0 sample. It can be seen from Fig.7 that the red area on the surface of the NC0 sample is obvious, indicating that the compressive stress causes great strain in the concrete at this phase. The red circle and black circle in the sample in Fig.7 show obvious cracking. At these crack locations, it is difficult to calculate the strain value using DIC, therefore, the color of these areas is white.

The same experiment was performed three times for the concrete prism that numbered NC0, RC25, RC50, RC75, RC100, and the experiment results were selected for comparison. The stress−strain curves of the experiment samples with different replacement rates of RCA and water−cement ratios are shown in Fig.7.

By comparing the stress−strain curves of samples with different mix proportions, it can be seen that the stress−strain curve of RAC was similar to NAC, and both consisted of an ascending section and a descending section. At the initial loading stage, the curve was close to a straight line, and the stress and strain conformed to a linear relationship. As loading was increased, it can be seen that the curve slope of the sample decreased with the increasing stress level. The curve of the sample began to decrease when the stress increased to the peak point, which is the axial compressive strength of the samples at the 28 d-curing age, and the corresponding strain at this point is considered the peak strain. It can be found from the stress−strain curve that the peak stress point of RAC samples was higher than that of the NC0 sample, and the peak stress point increased with the increase of the replacement rate of RCA. The specific reason for this phenomenon is that a certain amount of old mortar was attached to the surface of RCA. Compared with NCA, RCA has a lower strength, but a higher water absorption rate. In the process of mixed concrete, compared with the concrete with a low RCA replacement rate, the water in the concrete with a high RCA replacement rate was absorbed by more old mortar, and this absorption process was very fast. Therefore, RCA had already absorbed water before the concrete was fully hardened, which caused the actual water−cement ratio of the sample to be lower than the nominal water−cement ratio. In addition, the decrease of the water−cement ratio caused the strength of the concrete to increase. The peak strain of RAC samples was also higher than that of the NC0 sample, but there is no clear rule for the peak strain of different RAC samples. The following reasons may be caused this phenomenon. 1) The old mortar adhered to the surface of the RCA increased the gel content of the RAC after solidification, and the increase in the gel content increased the peak strain of RAC, which is a favorable factor. However, there were many microscopic cracks in the RCA itself. These cracks expanded when subjected to force, and thereby accelerated the destruction of concrete, which is an unfavorable factor. These two factors acted simultaneously during the stress and deformation process of RAC, and they cancelled each other out [15]. 2) The relatively higher water−cement ratio resulted in relatively lower gel content of the concrete after solidification, which resulted in relatively high porosity. Furthermore, the relatively higher porosity resulted in the overall performance of the experiment sample being weak and the peak strain being reduced.

Based on the above explanation, the experiment result was analyzed. It can be seen that in the RC25 sample, the effect of unfavorable factors was more evident because its peak strain is smaller than that of other RAC samples. The peak strain of the RC100 sample was greater than the RC25 sample because its water−cement ratio was significantly smaller than that of the RC25 sample, but its peak strain was smaller than those of the RC50 sample and the RC75 sample. The phenomenon may have occurred because the water−cement ratio of the RC50 sample and the RC75 sample is similar to the RC100 sample, and the RCA absorbed excess water, which made up for the internal defects of both of the RC50 sample and the RC75 sample, resulting in the decrease of the actual water−cement ratio and the increase of the peak strain. However, the RC100 sample had the lowest water−cement ratio and the highest replacement rate of RCA, so the residual water after cement hydration reaction was not enough to be absorbed by all microcracks in the RC100 sample, resulting in the residual water being compensating for only some of the internal defects of the RC100 sample. Therefore, the peak strain of the RC100 sample decreased. While the peak strain of the RC75 sample was slightly larger than the RC50 sample, indicating that the favorable factors were somewhat more evident in the RC75 sample. In addition, the NC0 sample had the highest water−cement ratio and no old mortar is attached, so it could not absorb excess water, which led to its lowest peak strain.

In the descending section after the peak point, the curve of RAC was steeper than that of NAC, indicating that RCA had more significant damage defects. These defects led to faster crack development in RAC than in NAC, and the brittleness of RAC is more evident than that of NAC.

Through the comparison between the stress−strain curves of different samples, it can be seen that before the peak point, the increase of the replacement rate of RCA could effectively improve the brittleness of RAC and increase its peak strain. After the peak point, there were natural defects in the RCA, which accelerated the damage of RAC. Therefore, the brittleness of concrete after the peak point becomes obvious with the increase of the replacement rate of RCA. Xiao et al. [35] found that decreasing the water−cement ratio also decreased the ductility of RAC. Based on the experiment result, the RC100 sample and the NC0 sample were selected for comparison because their water−cement ratio and replacement rate of RCA were very different, which can effectively explain the influence of the water−cement ratio and replacement rate of RCA on concrete brittleness. The strain corresponding to the stress value equal to 85% peak stress in the descending section of the stress−strain curve of the experiment sample was taken as the ultimate strain. The brittleness of different samples was compared by comparing the ratio of peak strain to the ultimate strain; the lower this ratio, the higher the brittleness of concrete. It can be seen that the peak strain of the NC0 sample was much lower than that of the RC100 sample, but the brittleness of the RC100 was more evident after the peak point. This phenomenon indicated that before the peak strain, the replacement rate of RCA could adjust the brittleness of concrete more effectively than the water−cement ratio. However, after the peak strain, the higher the replacement rate of RCA made the brittleness of concrete more evident.

3.2 Experimenting of mechanical properties of concrete beams

DIC provided the mid-span deflection of the concrete beam during the loading process, and the whole loading change process of the RC50 sample is displayed in Fig.8.

According to the load−deflection curve, the cracking deflection, the yield deflection and the ultimate deflection can be obtained, respectively. The results are shown in Tab.7.

Fig.8 shows that the mid-span deflection of the 5 concrete beams with different replacement rates of RCA all increased with the increase of the load value. The development trend of the RAC beam’s load−deflection curve was similar to that of the NC0 beam; both of them had three stages. The first stage was before concrete cracking; during the second stage concrete was cracking until the reinforcement yielded; during the third stage the reinforcement was yielding until the concrete became crushed. The following is the analysis of each stage.

In the first stage, the load was relatively small. The tensile strain of the concrete in tension zone was less than the ultimate tensile strain of the concrete. At this stage, the concrete did not crack yet. The slope of this stage gradually decreased with the increasing replacement rate of RCA. However, the slope of the RC75 sample was larger than that of the RC100 sample. This might be because the water−cement ratio of the RC75 sample was higher than that of the RC100 sample, resulting in the tensile strength of the RC75 sample being lower than that of the RC100 sample. It further shows that the influence of the water−cement ratio on concrete strength was more significant than that of the replacement rate of RCA. In the second stage, due to the low tensile strength of concrete, the tensile strain of concrete gradually increased to the ultimate value when the load was at about 10 kN, caused the concrete tensile zone to crack and the cracks began to expand. At this stage, the reinforcements bore the tensile stress at the concrete crack, resulting in the decrease of the slope of the curve of each sample. In addition, at the second stage, because the reinforcement was mainly in tension, and the longitudinal reinforcement of each experiment beam was the same, the slopes of each sample’s curve were almost the same. In the third stage, as the load continued to increase, the slope of the load−deflection curve decreased again, and an obvious inflection point appeared. This inflection point was the yield point of the reinforcement. After the inflection point, the load−deflection curve began to be smooth, and the deflection increased rapidly. The height of the compression zone of the experiment beam’s section gradually decreased. Finally, the third stage ended due to the destruction of the sample compression zone.

It can be seen from Tab.7 that the cracking deflection and yield deflection of the NC0 sample were generally both smaller than that of RAC sample. This may be due to the presence of micro-cracks in the RCA. These micro-cracks inhaled new cement particles, making the performance of ITZ between the mortar and the coarse aggregate was better than in the case of the NC0 sample. Except for the RC75 sample, the cracking deflection and yield deflection of RAC beams with different replacement rates of RCA were similar, with the difference being within 3%. However, abnormality occurred in the RC75 sample, the cracking deflection was too large, and the yield deflection was too small. The load−deflection curve of the samples suggests that except for the RC75 sample, the ultimate deflection value of the RAC samples was slightly larger than that of the NC0 sample, but the ultimate deflection of the RC75 sample was lower than for the NC0 samples. The abnormal experiment data of the RC75 sample might be due to the influence of the water−cement ratio, aggregate distribution and pouring quality. Furthermore, the cracks width of the RC75 sample was larger than those of other samples (this can be seen from Fig.9), this may also be related to the abnormal experimental data of the RC75 sample.

In addition, Tab.7 suggests that both the cracking deflection and yield deflection of RAC samples generally decreased with the increase of the RCA replacement rate, and also decreased with the decrease of the water−cement ratio. The reason was that the old mortar that remained attached to the surface of RCA had more microcracks, and these microcracks extended when being stressed and accelerated concrete cracking. It can be found from Tab.5 that in this experiment, the decrease of water−cement ratio meant the increase of cement content, and the decrease of residual water after the hydration reaction. This decrease led to the micro-cracks on the surface of RCA being unable to absorb excess residual water, which makes the brittleness of the RAC sample more significant. Because the experiment result of the RC75 sample is considered to be abnormal, the RC25 sample, the RC50 sample and the RC100 sample were selected for comparative analysis. It can be found that the difference of replacement rate of the RC25 sample and the RC50 sample is 25%, the water−cement ratio difference was 0.04, and the cracking deflection difference was 0.25 mm. The difference of the replacement rate of the RC50 sample and the RC100 sample was 50%, the water−cement ratio difference was 0.03, and the cracking deflection difference was only 0.02 mm. It shows that when the replacement rate of RCA was low, the water−cement ratio had a more significant impact on the cracking deflection of the RAC beams. The reason for this phenomenon is that when the replacement rate of RCA is low, the cracking of concrete was mainly determined by the properties of new mortar and natural aggregate. The yield deflections of the RC25 sample, the RC50 sample and the RC100 sample decreased in turn because their cracking deflections decreased in turn; the earlier the concrete cracks, the earlier the reinforcement entered the yield stage. There is no apparent law of the ultimate deflection among the three samples, indicating that both the replacement rate of RCA and water−cement ratio did not affect the ultimate deflection of the RAC sample.

It is seen from Tab.7 that the cracking load values of the different samples were similar, and the yield load values and the ultimate load values were also similar. It can be seen that the bearing capacity of samples had no obvious relationship with the replacement rate of RCA and the water−cement ratio; it was mainly related to the reinforcement of samples.

3.3 Time-dependent deformation performance experiment of concrete beams

3.3.1 Beam crack development nephogram

Fig.9 shows that the strain nephogram of samples with different replacement rates of RCA after loading were obtained through DIC. The strain and crack position of each sample after loading can be clearly seen in Fig.9. The meaning of green area and red area is as described in Section 3.1.2.

After loading, the red area on the sample surface indicated that the concrete had an obvious tensile strain, which meant the sample had cracked here. The red color depth represented the severity of cracking. Comparing the nephogram of different samples, it can be found that the water−cement ratio and the replacement rate of RCA seem did not change the crack development form of the samples, because these cracks are mainly distributed in the mid-span area, and all samples have about 10 locations of cracks, with the length of 3−6 cm and the width of 0.1–0.2 mm.

3.3.2 Fitting analysis and long-term deflection development of beams

The deflection values of each sample at different times were obtained by using the DIC. The deflection of the first day’s loading was defined as the initial deflection. The initial deflection of samples with different replacement rates of RCA is shown in Fig.10. It can be found that the initial deflection of the NC0 sample was about 20% higher than the initial deflection of other samples. Except for the RC50 sample, the initial deflection of other RAC samples increased with the increase of the replacement rate of RCA and the decrease of water−cement ratio. The small leap of the initial deflection of the RC50 sample may be because it’s located at the lowest layer of the loading device and bearing a more significant load.

The relative deflection value of each sample can be obtained by subtracting the initial deflection from the deflection every day. The mid-span relative deflection of the sample was selected for analysis, and the experiment data were curve fitted. Combined with the influence coefficient of the water−cement ratio proposed in Ref. [36], the relationship of relative deflection with experiment days obtained by curve fitting was as follows:

(1)\begin{array}{rl}f=& \alpha +\beta \times e^R\times \left(1-t^{\gamma }/\nu \right)\times (\epsilon \times (w/c{)}^2\\ & +\zeta \times (w/c)+\eta ),\left(1⩽t⩽394\mathrm{d}\right).\end{array}

In this equation, α represents the initial deflection, t represents the experiment days, R represents the replacement rate of RCA, and w/c represents the water−cement ratio. β, γ, ν, ε, ζ, and η represent the fitting coefficients. The values of each fitting coefficient are shown in Tab.8. The fitted relative deflection−time curve is shown in Fig.10. In Fig.10, the hollow symbol represents the measured value and the line represents the fitting curve.

Fig.10 shows that in the early loading phase (0–200 d), the value of the NC0 sample relative deflection was larger than that of the RC25 sample and the RC100 sample, and similar to the RC50 sample, but smaller than the RC75 sample. The relative deflection of the NC0 beams also increased faster. Experiment also shows that the deformation of the NC0 sample increased quickly in the early loading phase [32,37]. This may be because when the reinforcement of all samples was the same, the relative deflection of samples was affected by factors such as water−cement ratio, the replacement rate of RCA and load [36–38]. The increasing replacement rate of RCA means that there is an increase in the old mortar content in RAC, which promotes the growth rate of the samples’ relative deflection [7]. Reducing the water−cement ratio can reduce the relative deflection of samples [36]. The load also affects the relative deflection of the sample [31]. The relative deflection of the sample was greatly affected by the dominant factors, and the dominant factors of each sample were different. Therefore, the 0–200 days’ change of the relative deflection of RAC samples were not as evident as it was in the case of the NC0 sample. It can be seen that the relative deflection and increasing rate of the RC50 sample and the RC75 sample were also relatively large, the reason might be they were located in the lower layer of the loading device and bore a relatively large load, so it accelerated the development of micro-cracks in the RCA.

After 200 d, it can be seen that the relative deflection growth trend of each sample was the same. However, the relative deflection growth rate of the RAC sample was faster than the NC0 sample and became more evident with the increase of the replacement rate of RCA. This phenomenon shows that the replacement rate of RCA did not affect the tendency for the relative deflection of RAC samples to increase with time, but it affected the growth rate of relative deflection of RAC samples. The relative deflection value of the RC100 sample was the lowest of all samples at 0–200 d phase. However, it is not difficult to find from the development trend of the fitting curve after 200 d that the relative deflection of the RC100 sample increased the fastest. This phenomenon further shows that the replacement rate of RCA had an obvious effect on the growth rate of sample’s relative deflection.

Fig.10 illustrates that the relative deflection value and growth rate of RAC samples increased with the increase of the replacement rate of RCA except for the RC100 sample. The relative deflection value of the RC100 sample was the lowest of all samples, because of its minimum water−cement ratio, resulting in lower porosity, the hardened cement paste matrix was more uniform than the case for other samples, and the ITZ between the aggregate and the slurry has been improved. In addition, the RC100 sample bent upwards in the experiment, caused the relative deflection of the RC100 sample to grow slowly. The relative deflection of the RC25 sample was relatively low because of its lower replacement rate of RCA and fewer internal defects than other RAC samples. Besides, the RC25 sample was also upward bent and the growth of the relative deflection was less noticeable than the other downward bent samples. An interesting phenomenon was that although the replacement rate of the RC75 sample was greater than the RC50 sample, the RC75 sample’s relative deflection value was very close to the RC50 sample, the reason might be that the water−cement ratio and load of the RC75 sample were lower than the RC50 sample. This phenomenon shows that the water−cement ratio and load also affected the relative deflection value of samples. It is worth noting that the early growth rate of the RC75 sample’s relative deflection was the highest of all samples, which may be because the crack width of the RC75 sample after the end of loading was the largest of all samples, made the short-term stiffness of the RC75 sample the lowest of all samples.

Fig.10 exhibits that the influence on the relative deflection of the water−cement ratio on the early loading stage of the sample was more noticeable than the replacement rates of RCA because the relative deflection value of the RC100 sample was smaller than that of the NC0 sample, which is slightly different from the results of some previous experiments [39]. This phenomenon is due to the highest water−cement ratio of the NC0 sample, which led to the NC0 sample’s bearing capacity being the lowest of all samples, so the 0–200 d relative deflection of the NC0 sample developed relatively quickly. However, the replacement rate of RCA had a more significant impact on the relative deflection of samples at the later stage of loading, because after 200 d the growth rate of different RAC samples’ relative deflection became more evident with the increase of the replacement rate of RCA. This phenomenon was because the many micro-cracks in RCA will continue to crack and accelerate the relative deflection growth of RAC samples when the loading time reached a certain extent.

Fig.10 shows that the relative deflection growth trend of the RAC sample was similar to the NC0 sample, and the fitted relative deflection-time curve gradually flattened with the increase of time. Tab.9 suggests that Eq. (1) had a good fit for the long-term deformation of the samples, and the maximum error between the fitted value and the measured value was 3.28%. The fitting results were almost consistent with the measured data of this experiment and can be used as the reference for predicting the relative deflection of RAC beams.

However, it is worth noting that the change of relative deflection of the beam is very complex. It is not only affected by the replacement rate of RCA and water−cement ratio but also affected by drying shrinkage, icons attack, and other factors [40–42]. It is not clear how the coupling of these factors will affect the relative deflection of concrete beams. Therefore, the change of relative deflection of concrete beams under complex conditions and the generality of the relative deflection prediction equation proposed in this paper still needs to be further studied.

4 Conclusions

Compared with the traditional method, DIC can better analyze the entire crack growth process from initiation to propagation and the final penetration of the concrete components during the loading process, while the traditional method can only observe the apparent cracks and cannot know the crack development process and formation mechanism. In this paper, by changing the water−cement ratio and the replacement rate of RCA, axial compressive strength of concrete prism and short-term mechanical properties of concrete beams is studied. In addition, through the self-designed concrete beam loading device under constant loading and the DIC cameras, the long-term deformation properties of concrete beams with different replacement rates of RCA and water−cement ratios were measured. The following conclusions were drawn.

1) The axial compressive strength of concrete gradually increases with the decrease of the water−cement ratio and the replacement rate of RCA. The influence of the water−cement ratio on the strength change is more evident than the replacement rate of RCA. The strength of a low water−cement ratio sample can even be 70% higher than that of a high water−cement ratio sample.

2) Before the peak strain, the replacement rate of RCA can adjust the brittleness of concrete more effectively than the water−cement ratio. After the peak strain, the higher the replacement rate of RCA will make the brittleness of concrete more obvious.

3) The ultimate bearing capacity of concrete beams has little relation to the replacement rate of RCA and the water−cement ratio and is mainly related to their reinforcement. After adjusting the mix proportion, the ultimate strength of the RAC beams is 15% higher than that of the NAC beams.

4) The cracking deflection and yield deflection of RAC beams decrease with the increase of the replacement rate of RCA. When the replacement rate of RCA is low, the water−cement ratio has a more significant impact on the deflection of RAC beams. The yield deflection and ultimate deflection mainly depend on the reinforcement and have little relationship with the replacement rate of RCA and water−cement ratio.

5) The relative deflection value and growth rate of RAC beams generally increase with the increase of the replacement rate of RCA. However, the water−cement ratio and load will also have different effects on the relative deflection value. The water−cement ratio has a more pronounced influence on the relative deflection of the concrete beam at the early stage of loading, and the replacement rate of RCA has a more significant influence on the relative deflection of the concrete beam at the later stage of loading. However, the change of relative deflection of the concrete beam is very complex, which will be affected by many factors.

6) The long-term deformation results of curve fitting are in good agreement with the measured values. The maximum error is 3.28%, which can be used as the reference for the long-term deformation prediction of behaviour of RAC beams.

References

[1] Xiao J, Li W, Poon C. Recent studies on mechanical properties of recycled aggregate concrete in China—A review. Science China Technological Sciences, 2012, 55( 6): 1463–1480

[2] de Brito J, Ferreira J, Pacheco J, Soares D, Guerreiro M. Structural, material, mechanical and durability properties and behaviour of recycled aggregates concrete. Journal of Building Engineering, 2016, 6 : 1–16

[3] Casuccio M, Torrijos M C, Giaccio G, Zerbino R. Failure mechanism of recycled aggregate concrete. Construction & Building Materials, 2008, 22( 7): 1500–1506

[4] Chindaprasirt P. Handbook of Alkali-Activated Cements, Mortars and Concretes. Sawston, Cambridge: Woodhead Publishing, 2015, 519–538

[5] Nagataki S, Gokce A, Saeki T, Hisada M. Assessment of recycling process induced damage sensitivity of recycled concrete aggregates. Cement and Concrete Research, 2004, 34( 6): 965–971

[6] Khatib J M. Properties of concrete incorporating fine recycled aggregate. Cement and Concrete Research, 2005, 35( 4): 763–769

[7] Fan Y, Xiao J, Tam V W Y. Effect of old attached mortar on the creep of recycled aggregate concrete. Structural Concrete, 2014, 15( 2): 169–178

[8] Domingo-Cabo A, Lázaro C, López-Gayarre F, Serrano-López M A, Serna P, Castaño-Tabares J O. Creep and shrinkage of recycled aggregate concrete. Construction & Building Materials, 2009, 23( 7): 2545–2553

[9] Topçu I B, Günçan N F. Using waste concrete as aggregate. Cement and Concrete Research, 1995, 25( 7): 1385–1390

[10] Tocpu I B. Physical and mechanical properties of concretes produced with waste concrete. Cement and Concrete Research, 1997, 27( 12): 1817–1823

[11] Sato R, Maruyama I, Sogebe T, Sogo M. Flexural behavior of reinforced recycled concrete beams. Journal of Advanced Concrete Technology, 2007, 5( 1): 43–61

[12] Ma H, Xue J, Zhang X, Luo D. Seismic performance of steel-reinforced recycled concrete columns under low cyclic loads. Construction & Building Materials, 2013, 48 : 229–237

[13] Corinaldesi V, Letelier V, Moriconi G. Behaviour of beam−column joints made of recycled-aggregate concrete under cyclic loading. Construction & Building Materials, 2011, 25( 4): 1877–1882

[14] XiaoJLiJSunZHaoX. Study on compressive strength of recycled aggregate concrete. Journal of Tongji University, 2004, 32(12): 1558–1561 (in Chinese)

[15] ChenZXuJZhengHSuYXueJLiJ. Basic mechanical properties test and stress−strain constitutive relation of recycled aggregate concrete. Journal of Building Materials, 2013, 16(1): 24–32 (in Chinese)

[16] Ignjatović I S, Marinković S B, Mišković Z M, Savić A R. Flexural behavior of reinforced recycled aggregate concrete beams under short-term loading. Materials and Structures, 2013, 46( 6): 1045–1059

[17] Etxeberria M, Marí A R, Vázquez E. Recycled aggregate concrete as structural material. Materials and Structures, 2007, 40( 5): 529–541

[18] Adam M, Knaack Y C K. Behavior of reinforced concrete beams with recycled concrete coarse aggregates. Journal of Structural Engineering, 2015, 141( 3): B4014009

[19] Choi W, Yun H. Long-term deflection and flexural behavior of reinforced concrete beams with recycled aggregate. Materials & Design, 2013, 51 : 742–750

[20] Arezoumandi M, Smith A, Volz J S, Khayat K H. An experimental study on flexural strength of reinforced concrete beams with 100% recycled concrete aggregate. Engineering Structures, 2015, 88 : 154–162

[21] YangJWuJYeQ. Study on deflection of recycled concrete beams. Engineering Mechanics, 2011, 28(2): 147–151 (in Chinese)

[22] GB50010-2010(2015 Edition). Code for Design of Concrete Structures. Beijing: China Academy of Building Sciences, 2016 (in Chinese)

[23] MaruyamaISogoMSogabeTSatoRKawaiK. Flexural properties of reinforced recycled concrete beams. In: Proceedings of the International RILEM Conference on the Use of Recycled Materials in Buildings and Structures. Barcelona: RILEM, 2004

[24] Aggelis D G, Verbruggen S, Tsangouri E, Tysmans T, van Hemelrijck D. Characterization of mechanical performance of concrete beams with external reinforcement by acoustic emission and digital image correlation. Construction & Building Materials, 2013, 47 : 1037–1045

[25] Alam S Y, Saliba J, Loukili A. Fracture examination in concrete through combined digital image correlation and acoustic emission techniques. Construction & Building Materials, 2014, 69 : 232–242

[26] Destrebecq J F, Toussaint E, Ferrier E. Analysis of cracks and deformations in a full scale reinforced concrete beam using a digital image correlation technique. Experimental Mechanics, 2011, 51( 6): 879–890

[27] Huang Y, He X, Wang Q, Xiao J. Deformation field and crack analyses of concrete using digital image correlation method. Frontiers of Structural and Civil Engineering, 2019, 13( 5): 1183–1199

[28] Li W, Xiao J, Sun Z, Shah S P. Failure processes of modeled recycled aggregate concrete under uniaxial compression. Cement and Concrete Composites, 2012, 34( 10): 1149–1158

[29] Gencturk B, Hossain K, Kapadia A, Labib E, Mo Y L. Use of digital image correlation technique in full-scale testing of prestressed concrete structures. Measurement, 2014, 47 : 505–515

[30] CaoWXiaoJYeTLiL. Research progress and engineering application of reinforced recycled aggregate concrete structure. Journal of Building Structures, 2020, 41(12): 1–16 (in Chinese)

[31] CaoWPengSQiaoQLinDZhuK. Experimental study on deformation performance of full-sized high strength recycled reinforced concrete beams under long-term loading. Journal of Building Structures, 2017, 38(11): 142–148 (in Chinese)

[32] LiuCBaiGZhangYYinY. Study on stiffness calculation method and flexural performance of recycled concrete beam based on long-term loading effect. Journal of Building Structures, 2016, 37(S2): 1–8 (in Chinese)

[33] GuoYLiQYueGLiQ. Calculation of compressive strength of recycled concrete based on coarse aggregate quality and replacement rate. Journal of Building Structures, 2018, 39(4): 153–159 (in Chinese)

[34] LianHLiY. Discussion on method for selecting mix proportion of concrete (II): Proposal on selection of factors and method for calculating mix proportion of current concrete. Concrete (London), 2009, 31(5): 1–4 (in Chinese)

[35] Xiao J, Li W, Sun Z, Shah S P. Crack propagation in recycled aggregate concrete under uniaxial compressive loading. ACI Materials Journal, 2012, 109( 4): 451–461

[36] ZhaoMYangHWangYGengY. Creep model for recycled coarse and fine aggregate concrete considering water−cement ratio of matrix concrete. Journal of Building Structure, 2020, 41(12): 148–155 (in Chinese)

[37] LiuCZhuCFanJBaiGXiaoJ. Experimental study on recycled concrete beam subjected to long-term loading and its flexural performance after unloading. Journal of Building Structures, 2020, 41(12): 56–63 (in Chinese)

[38] Geng Y, Wang Y, Chen J. Creep behaviour of concrete using recycled coarse aggregates obtained from source concrete with different strengths. Construction & Building Materials, 2016, 128 : 199–213

[39] Domingo A, Lázaro C, Gayarre F L, Serrano M A, López-Colina C. Long term deformations by creep and shrinkage in recycled aggregate concrete. Materials and Structures, 2010, 43( 8): 1147–1160

[40] Sun D, Shi H, Wu K, Miramini S, Li B, Zhang L. Influence of aggregate surface treatment on corrosion resistance of cement composite under chloride attack. Construction & Building Materials, 2020, 248 : 118636

[41] Sun D, Wu K, Shi H, Miramini S, Zhang L. Deformation behaviour of concrete materials under the sulfate attack. Construction & Building Materials, 2019, 210 : 232–241

[42] Koh C G, Ang K K, Zhang L. Effects of repeated loading on creep deflection of reinforced concrete beams. Engineering Structures, 1997, 19( 1): 2–18

RIGHTS & PERMISSIONS

Higher Education Press