Acoustic emissions evaluation of the dynamic splitting tensile properties of steel fiber reinforced concrete under freeze–thaw cycling

Abstract

This study empirically investigated the influence of freeze–thaw cycling on the dynamic splitting tensile properties of steel fiber reinforced concrete (SFRC). Brazilian disc splitting tests were conducted using four loading rates (0.002, 0.02, 0.2, and 2 mm/s) on specimens with four steel fiber contents (0%, 0.6%, 1.2%, and 1.8%) subjected to 0 and 50 freeze–thaw cycles. The dynamic splitting tensile damage characteristics were evaluated using acoustic emission (AE) parameter analysis and Fourier transform spectral analysis. The results quantified using the freeze–thaw damage factor defined in this paper indicate that the degree of damage to SFRC caused by freeze–thaw cycling was aggravated with increasing loading rate but mitigated by increasing fiber content. The percentage of low-frequency AE signals produced by the SFRC specimens during loading decreased with increasing loading rate, whereas that of high-frequency AE signals increased. Freeze–thaw action had little effect on the crack types observed during the early and middle stages of the loading process; however, the primary crack type observed during the later stage of loading changed from shear to tensile after the SFRC specimens were subjected to freeze–thaw cycling. Notably, the results of this study indicate that the freeze–thaw damage to SFRC reduces AE signal activity at low frequencies.

Graphical abstract

Keywords

steel fiber reinforced concrete / freeze–thaw cycling / Brazilian disc splitting test / acoustic emission technique / dynamic splitting tensile acoustic emission properties

1 Introduction

Concrete has been widely used around the world owing to its excellent performance; however, its low tensile strength represents a critical vulnerability that limits its further application in practical projects. Moreover, freeze–thaw damage is a common issue for concrete infrastructure in cold areas as it causes quality problems such as volume expansion, surface loss, and cracking that adversely impact the durability and integrity of structures and often result in considerable economic losses. Wang et al. [1] verified an existing analytical model characterizing the two-way coupling effect of chloride ion migration and water diffusion in concrete by establishing the initial and boundary conditions considering the actual environment, which was beneficial for the predicting the service life of concrete structures. Liu et al. [2] conducted accelerated corrosion tests on reinforced concrete samples under freeze–thaw cycling and used digital image correlation technology to survey their surface strains. Wang et al. [3] studied the relationship between freeze–thaw damage and pore structure degradation using low-field nuclear magnetic and mercury intrusion porosimetry and established pore structure deterioration equations to estimate the remaining life of mortar subjected freeze–thaw cycling. To better protect reinforced concrete highway bridge structures from chemical attack and particularly exposure to freeze–thaw conditions, the Colorado Department of Transportation laid waterproof membranes over a bridge, thereby relieving deterioration [4]. The addition of steel fibers to concrete cannot only enhance its tensile strength, but also improve its freeze–thaw cycle resistance [5–9]. Nili et al. [10] demonstrated that minimum polypropylene and steel fiber volumes of 0.4% and 0.5%, respectively, were required to realize improved freeze–thaw durability. Indeed, steel fiber reinforced concrete (SFRC) has become the most frequently used fiber reinforced concrete as it is well-suited for use in cold regions to reduce structural damage owing to freeze–thaw cycling.

Many studies investigated the mechanical properties of SFRC subjected to freeze–thaw cycling. Dong et al. [11] conducted a series of tests demonstrating that SFRC exhibited superior compressive strength compared to polypropylene fiber reinforced concrete and proved that a reduced interfacial coherence between the fiber and matrix was the primary cause of deformability deterioration. Rapid freeze–thaw tests were performed on steel fiber coal gangue concrete by Qiu et al. [12], who concluded that the mass loss and relative dynamic elastic modulus of coal gangue concrete can be reduced by adding an appropriate quantity of steel fibers. Furthermore, the impacts of freeze–thaw cycling and the addition of steel fibers on the stress–strain curve and ultimate axial compressive strength of SFRC were studied by Zhou et al. [13]. However, previous studies subjecting SFRC to freeze–thaw cycling have primarily focused on the compressive strength and elastic modulus of specimens while neglecting their tensile properties, especially under dynamic loads. Furthermore, the mechanical properties of SFRC are typically investigated using macro tests, which cannot provide a detailed explanation of the damage characteristics owing to freeze–thaw cycling during the SFRC failure process. Consequently, dynamic tensile testing of SFRC is necessary to investigate its damage features when subjected to freeze–thaw cycling.

The acoustic emission (AE) technique is a dynamic and nondestructive approach for evaluating, locating, and forecasting concrete damage. Both parametric and waveform analyses play essential roles in analyzing the AE signals produced by concrete materials. Parametric analysis is a classical method that has the advantages of simplicity and speed. The variation and distribution of AE parameters, including the signal frequency, amplitude, and number of AE events, can indirectly reflect the microscopic damage process inside concrete [14–16]. Yue et al. [17] proposed a new method to identify the tension and shear crack modes based on the analytical peak frequency of AE signals. Zhang et al. [18] applied an AE parametric analysis to investigate the dynamic splitting performance of reinforced concrete with basalt fibers and stated that both the loading rate and basalt fiber dosage played significant roles. Qiu et al. [19] defined a splitting damage variable for coal gangue concrete with steel fibers and established a failure model according to fiber content based on the AE energy. Waveform analysis has been increasingly employed in recent years. Fourier transform (FT) analysis is conducive to obtaining the frequency-domain information of AE signals, which cannot be obtained by parametric analysis. Lai et al. [20] conducted AE tests of concrete specimens subjected to uniaxial compression and analyzed the original waveform information using the FT method to study the damage mechanism. Topolar et al. [21] analyzed the AE signals of air-entrained concrete using a fast Fourier transform (FFT) to assess the resistance of concrete to freeze–thaw action; they concluded that the dominant frequency of concrete decreases during freeze–thaw cycling. Both the AE parameter and waveform analysis methods represent mature technologies with considerable potential for studying the failure of SFRC subjected to freeze–thaw cycling. However, few studies have investigated the dynamic tensile damage performance of SFRC under freeze–thaw cycling using an AE system.

This study accordingly investigated the dynamic splitting tensile damage performance of SFRC under freeze–thaw cycling using the AE technique. A series of Brazilian disc splitting tests was performed using the loading rate, fiber content, and application of freeze–thaw cycles as variables to investigate the splitting tensile damage characteristics of SFRC through AE parametric and FT analyses. The results of this study can serve to guide further study of damage monitoring systems for SFRC structures subjected to freeze–thaw cycling.

2 Experiment program

2.1 Preparation of specimens

As the fiber contents of the SFRC specimens considered in previous studies were primarily concentrated within 0%–2% [22–24], four different fiber contents (0%, 0.6%, 1.2%, and 1.8%) were evaluated in the dynamic splitting tensile tests conducted in this study. The specimens prepared in this study comprised 42.5-grade ordinary Portland cement with river sand and natural gravel as the fine and coarse aggregates, respectively. Tab.1 lists the mix proportions of these specimens [18]. Hooked-end steel fibers, depicted in Fig.1 with the physical and mechanical properties listed in Tab.2, were dry mixed with the gravel for 1 min before the other raw materials were added to ensure an even distribution of fibers in the concrete. The fully mixed concrete was subsequently poured into \mathrm{\Phi }100 mm × 50 mm Brazilian disc molds. Form stripping was performed after 48 h of initial curing, followed by 28 d of curing in standard conditions at a temperature of (20 ± 2) °C and relative humidity of 95%. A cured specimen is shown in Fig.2. The freeze–thaw cycling experiments were performed using an HDK rapid freeze–thaw testing machine. A rapid freezing method was applied, with the machine set to perform 25 cycles within 6 d. The specific conditions of this test are detailed in Tab.3. Three sets of parallel tests were conducted for each working condition, and the splitting tensile strength was calculated as the average of these three tests.

2.2 Brazilian disc splitting tests

The splitting tensile strength of SFRC was evaluated in this study using the Brazilian disc test. The theoretical equation for the splitting tensile strength of SFRC was derived based on elastic theory and the Griffith strength criterion and can be expressed as follows [25,26]:

(1)f_{\mathrm{t}}=\frac{2P_{max}}{\text{π}{D}_{\mathrm{s}}H},

where f_{\mathrm{t}} is the splitting tensile strength of the specimen, D_{\mathrm{s}} is the diameter of the disc specimen, P_{max} is the peak load, and H is the height of the specimen.

An Instron 8802 electro-hydraulic servo fatigue testing machine, illustrated in Fig.3, was used to perform the Brazilian disc splitting tests. Cork pads were placed on the upper and lower ends of the Brazilian disc specimens to relieve stress concentrations. A monotonic compression load was applied as the test loading, and the testing machine stopped loading when the specimen was completely broken. As an AE analysis requires a significant number of AE signals, the default loading rate applied by the testing machine could be too fast to collect sufficient AE signals; therefore, four loading rates (0.002, 0.02, 0.2, and 2 mm/s) were evaluated for each set of test parameters.

2.3 AE test

A photograph of the experimental setup comprising the Brazilian disc splitting test system and the AE detection equipment is shown in Fig.4. A PCI-2 AE system was employed using the setup shown by the schematic in Fig.5. An AE is a transient elastic wave caused by elastic deformation and cracking inside a material. Surface displacements can be detected by AE sensors when AE signals propagate to the surface of the specimen. The sensor converts this surface displacement and mechanical vibration into an electrical signal to obtain AE signal parameters and waveform data. Tab.4 lists the AE parameters that were set according to the material used in this test and the relevant environmental noise. Parameter and fast FT analyses were subsequently performed on the collected AE data.

The amplitude, ring count, energy count, rise time, frequency, and other parameters of the AE signal, illustrated in Fig.6, were used to evaluate the concrete damage level and identify the crack type. Shahidan et al. [27] used AE parameters such as amplitude, rise time, and signal strength to categorize the damage to and identify the damage level of a reinforced concrete beam. In this study, the energy count, peak frequency, rise angle, and average frequency parameters were analyzed to study the dynamic mechanical damage characteristics of SFRC subjected to freeze–thaw cycles. A large quantity of hidden AE information can be additionally obtained through frequency-domain analysis. The FT can convert the time–amplitude curve of the AE signal into a frequency–amplitude graph. Currently, the FT is typically performed on computers as a discrete Fourier transform (DFT) [28]. The continuous FT of the discrete-time signal x\left(n\right) is defined as:

(2)x\left({\text{e}}^{\mathrm{j}w}\right)= \sum _{n=-\mathrm{\infty }}^{\mathrm{\infty }}x\left(n\right) {\text{e}}^{-\mathrm{j}wn},

where x\left({\text{e}}^{\mathrm{j}w}\right) is a continuous function, {\text{e}}^{\mathrm{j}wn}=\cos \left(wn\right)+ \sin \left(wn\right) and the x\left(n\right) spectrum must be discretely approximated. The DFT of a finite discrete signal x\left(n\right) is defined as

(3)X\left(k\right) =\sum _{n=0}^{N-1} x\left(n\right)W_N^{nk},

where W_N={\text{e}}^{-\mathrm{j}\frac{2\text{π}}{N}} and k = 0,1,...,N-1. The inverse transformation of Eq. (3) is defined as:

(4)x\left(n\right)=\frac{1}{N}\sum _{k=0}^{N-1} X\left(k\right)W_N^{-nk}.

According to the definition of the DFT, x\left(n\right) must be multiplied N times and added \left(N-1\right) times for each n. Thus, a total of N^2 complex multiplications and N\left(N-1\right) complex additions are required for an n-point transformation. As a result, a significant amount of time is required to perform a DFT when N is large.

Cooley and Tukey [29] proposed the FFT in 1965 to solve the problem of complex and time-consuming calculation. The FFT takes advantage of the periodicity and symmetry of the W_N factor to reduce the number of operations N^2 to \left(N/2\right) {\log }_{2}N, providing a revolutionary method for processing digital signals under the premise of ensuring timely DFT calculations.

3 Results and discussions

3.1 Analysis of dynamic splitting tensile strength

Fig.7(a) and Fig.7(b) describe the changes in the dynamic splitting tensile strength of SFRC specimens with different fiber contents subjected to different loading rates after 0 and 50 freeze–thaw cycles, respectively. The splitting tensile strength of the SFRC specimens increased with the loading rate as well as the fiber content. As shown in Fig.7(a), the growth rate of splitting tensile strength slowed and there was even a slight decrease in strength when the fiber content was greater than 1.2%. Specimens with fiber contents of 1.2% and 1.8% exhibited relatively superior splitting tensile strengths compared to those with other fiber contents because an appropriate quantity of steel fibers can form a mesh skeleton within the concrete to prevent the development of cracks. However, an excessive quantity of fibers will cause more cement to cling to the surfaces of the fibers, reducing the quantity of cement participating in the hydration reaction. In addition, excess steel fibers easily agglomerate. Indeed, an even distribution of fibers in a specimen is difficult to ensure; as a result, as the quantity of fibers increases, the distribution of fibers inside the concrete becomes more uneven, producing more areas of local weakness. Therefore, a high fiber content cannot further increase the dynamic splitting tensile strength of SFRC.

As illustrated in Fig.7(b), a decrease in the splitting tensile strength of the SFRC specimens was observed after 50 freeze–thaw cycles. The strain rate effect on the specimen was significantly reduced after freeze–thaw cycling, and its growth rate decreased. Original damage, such as microcracks and micropores, was randomly distributed inside the SFRC material during preparation and aggravated to different extents after freeze–thaw cycling. Clearly, any moisture inside the SFRC will be frozen under freeze–thaw conditions, resulting in volumetric expansion that applies squeezing frost heave pressure to the pores, leading to the gradual expansion and connection of pores in the specimen and thereby aggravating extant concrete damage. Indeed, the concrete specimens were more damaged and their integrity reduced after freeze–thaw cycling. Fig.7(b) shows a linearly increasing relationship between the splitting tensile strength and steel fiber content. Notably, the optimal of steel fiber content for SFRC subjected to freeze–thaw cycling was 1.8%, which is different from the results presented in Fig.7(a). Thus, the results indicate that SFRC with different quantities of steel fibers will suffer different degrees of damage after 50 freeze–thaw cycles, leading to different degrees of decrease in the splitting tensile strength.

To further investigate the impact of freeze–thaw cycling on the dynamic mechanical properties of SFRC, the freeze–thaw damage factor D was defined as follows:

(5)D=\frac{{\sigma }_0-{\sigma }_{50}}{{\sigma }_0},

where {\sigma }_0 and {\sigma }_{50} are the splitting tensile strengths of SFRC subjected to 0 and 50 freeze–thaw cycles, respectively.

The D values for the various SFRC specimens are plotted in Fig.8 according to testing condition. Clearly, D decreased as the steel fiber quantity increased (except at 2 mm/s), indicating that the frost resistance of SFRC increased with the addition of more end-hook-type steel fibers. However, the influence of the loading rate on the freeze–thaw damage factor of SFRC exhibited no consistent law.

3.2 Analyses of AE parameters

3.2.1 Energy count analysis

The energy count is calculated as the area under the envelope of the signal waveform and reflects the relative energy or intensity of an event, as shown in Fig.6. The results for a loading rate of 0.02 mm/s were employed to investigate the impact of steel fiber content on the energy count. Fig.9 and Fig.10 plot the energy count of the SFRC specimens according to steel fiber content when subjected to 0 and 50 freeze–thaw cycles, respectively. The figures indicate that the fiber content strongly affected the energy count during the splitting tensile destruction of the SFRC specimens. The high-energy AE signals were primarily concentrated in the later stage of loading for plain concrete, whereas the high-energy AE signals generated by the SFRC specimens were distributed throughout the entire test. This indicates that the AE signals produced by SFRC during splitting tensile damage are released steadily when the appropriate fiber content is provided. High-strength signals were released suddenly when the plain concrete was destroyed, indicating that the plain concrete experienced a brittle tensile splitting failure. However, the development of concrete cracking can be restrained by the bridging action of the steel fibers, though this can also adversely impact the ductility of the concrete. The total energy count of the AE signals produced by the specimens increased with increasing fiber content owing to the enhancement of the resistance provided against external forces by the incorporation of steel fibers.

Comparing Fig.9 and Fig.10, the energy count and total accumulated energy count decreased to a certain degree after freeze–thaw cycling, demonstrating that the splitting tensile strength was weakened by frost action. Indeed, the coherence between the concrete matrix and steel fibers is poor when concrete is subjected to freeze–thaw cycling. However, freeze–thaw cycling generated a larger number of small pores within the concrete, improving the capacity of the SFRC to absorb energy. Therefore, the energy released during the failure of the specimens exposed to freeze–thaw cycling was lower than that of the specimens that were not exposed to freeze–thaw cycling. In addition, Fig.9(a) and Fig.10(a) illustrate that the energy count of the AE signals produced by the plain concrete decreased considerably after freeze–thaw cycling, indicating that the damage caused by frost action has a significant influence on plain concrete. In contrast, the energy count of the AE signals produced by the SFRC decreased only slightly after freeze–thaw cycling, demonstrating that the addition of steel fibers enhances the resistance of concrete to freeze–thaw cycling.

3.2.2 Peak frequency analysis

The energy count reflects the AE signal activity of the specimen during splitting tensile damage; however, it cannot be used to determine the sources of the AE signals. Therefore, the distribution of the peak AE signal frequencies produced by the various Brazilian disc specimens subjected to 0 and 50 freeze–thaw cycles were analyzed to investigate the AE signal sources. All AE signals exhibited a peak frequency distribution ranging from 0 to 300 kHz and could be separated into 12 small-spectrum bands each spanning 25 kHz. The ratios of the number of AE event hits in each small band to the total number of hits were subsequently calculated as plotted in Fig.11.

The AE peak frequency distribution was strongly influenced by the loading rate. Zhang et al. [18] demonstrated that the splitting of the concrete matrix generates a low-frequency AE signal below 100 kHz, and the fracture of the aggregate generates a high-frequency AE signal ranging from 100 to 200 kHz. Moreover, as the boundary between the mortar and aggregate represents the weak zone of the SFRC, this is where destruction begins. Only one peak frequency interval [0, 25] can be observed in the peak frequency distribution graph for the specimens subjected to a loading rate of 0.002 mm/s, indicating that splitting tensile damage cracks were generated in the Brazilian discs and extended along the boundary between the mortar and aggregate at lower loading rates. In contrast, two peak intervals, [0, 25] and [125, 175], can be observed in the peak frequency distribution graph for the specimens subjected to a loading rate of 0.02 mm/s. Two peak intervals can also be observed in the graphs for the relatively faster loading rates of 0.2 and 2 mm/s, concentrated in the intervals [0, 75] and [125, 175], respectively; note that the proportions of these two intervals are similar. More cracks extended through the aggregates at higher loading rates than that at lower loading rates because an rapid loading rate does not provide sufficient time for either original or new cracks inside the concrete to expand. Therefore, new cracks are readily generated owing to the increased stress in the surrounding concrete mortar matrix. As the loading rate continues to increase, stronger aggregates become more involved in the destruction of the concrete, representing one of the reasons for the observed increase in the dynamic splitting tensile strength of the SFRC specimens.

As plotted in Fig.11, the peak frequency distributions of plain concrete subjected to 50 freeze–thaw cycles exhibit approximately similar patterns. When the loading rate was 2 mm/s, one of the peak frequency intervals moved from (50, 75] to (25, 50] after freeze–thaw cycling because concrete that has not been subjected to freeze–thaw cycling has a higher-strength interface transition zone, whereas concrete subjected to freeze–thaw cycling exhibits an increased internal microporosity, causing a decrease in concrete strength. The higher the splitting tensile strength, the higher the corresponding frequency spectrum during damage, resulting in a change in the peak frequency interval. The interval containing the peak point did not change significantly when the loading rate was relatively low (0.002, 0.02, or 0.2 mm/s).

Fig.12(a) shows that the peak frequency distributions of the SFRC specimens were generally concentrated in two intervals: [0, 100] and [100, 200]. Research has shown that fiber splitting and tension tend to produce higher-frequency AE signals than matrix cracking [30,31]. The proportion of low-frequency band signals continued to decrease and the percentage of high-frequency band signals continued to increase as the loading rate increased, indicating that the slip and pull of the fibers generated high-frequency signals. Fig.12(b) indicates that the peak frequency distributions of the SFRC specimens subjected to 50 freeze–thaw cycles were generally more uniform compared to those subjected to 0 freeze–thaw cycles. Furthermore, the peak frequency distribution of plain concrete subjected to 50 freeze–thaw cycles exhibited poor distribution. Thus, the proportion of high-frequency signals produced by SFRC likely increased after 50 freeze–thaw cycles because the cohesion between the steel fibers and the matrix weakened, making it easier for the fibers to slip and pull out during loading.

Fig.13 shows the effects of fiber content on the peak frequency distributions produced by the SFRC specimens loaded at 0.02 mm/s. Because the loading rate was relatively slow, the proportion of low-frequency bands was considerable regardless of fiber content. As the steel fiber content increased, the proportion of low frequencies gradually diminished and the proportion of high frequencies gradually increased, indicating that a higher fiber content increases the prominence of high-frequency AE signals. The peak frequency proportions for the specimens subjected to 0 and 50 freeze–thaw cycles exhibited similar regularity, but the low-frequency proportion slightly decreased and the high-frequency proportion increased after freeze–thaw cycling.

3.2.3 Crack type analysis

The rise angle (RA) and average frequency (AF) represent the ratios of the rise time T_r to the amplitude A and the ring count n to the duration T_d, respectively, and are expressed as [32]:

(6)RA=T_{\mathrm{r}} /A ,

(7)AF=n/T_{\mathrm{d}} .

The relative relationship between RA and AF can be used to identify the crack types generated in concrete [18,33–35]. Tensile cracks with p-waves and shear cracks with s-waves are produced during concrete failure. As the propagation velocity of a p-wave is higher than that of an s-wave, the rise time of an AE signal generated by a p-wave is shorter. Therefore, the AE signals released when shear cracks occur exhibit higher RA and lower AF values [36].

Fig.14 and Fig.15 show the changes in the RA and AF values of the SFRC specimens with different fiber contents when subjected to 0 and 50 freeze–thaw cycles, respectively, and tested at a loading rate of 0.02 mm/s. The value of each point on the curve represents the average value calculated considering 30 proximate points. The RA and AF values exhibited similar change trends over time regardless of freeze–thaw cycling; these trends can be divided into three stages, as illustrated in Fig.14 and Fig.15. The RA values remained relatively small and the AF values were quite high during the first stage, indicating that this stage was dominated by tensile cracking. Subsequently, the RA values increased slightly to approach those of the AF in the second stage, though the latter remained higher, indicating that more tensile cracks were generated. Notably, the AF values for the SFRC specimens subjected to 50 freeze–thaw cycles did not exhibit as sudden a decrease as those for the equivalent specimens subjected to 0 freeze–thaw cycles during the second stage, but rather decreased with fluctuation, indicating that numerous tensile cracks were still generated during this period. This fluctuating decrease was likely caused by the larger quantity of initial pores in concrete that has been subjected to freeze–thaw cycling, leading to the appearance of numerous tensile cracks. There are no other significant differences between Fig.14 and Fig.15 in the first two stages of loading, suggesting that freeze–thaw cycling has little influence on the crack types produced in SFRC during the early and middle stages of tensile load. During the third loading stage, the RA and AF values were quite similar for the plain concrete, as shown in Fig.14(a). However, the difference between RA and AF was more distinct for specimens with higher fiber contents, indicating that the presence of steel fibers increased the proportion of shear cracks in the concrete during the final loading stage owing to their enhancement of its resistance to tensile forces. Additionally, Fig.15 shows that tensile cracks dominated during the later loading stage, except for Fig.15(c) (in which the specimens were subjected to insufficient freeze–thaw action). This behavior may be a result of matrix strength weaking owing to freeze–thaw cycling, preventing the steel fibers from playing a role in the tensile response before the concrete fails.

3.3 An FT analysis of AE signals

An FFT of the AE signals for each specimen was performed to analyze the waveform signals in the frequency domain. The SFRC with 1.2% fiber content was chosen for this analysis to observe the impact of loading rate on the spectral characteristics of its AE signal.

The primary frequencies in the spectra according to loading rate were concentrated around 20, 50, and 150 kHz, as plotted in Fig.16 and Fig.17. The energy of the AE signal was primarily released in the 40–170 kHz range. The AE signal at lower frequencies exhibited stronger energy with increasing loading rate, as the higher loading rates did not allow sufficient time for existing and new cracks inside the concrete to expand. Thus, the stress in the concrete mortar matrix around the cracks increased and new cracks were easily generated. Previous research has shown that the development of large-scale cracks tends to release relatively low-frequency AE signals [37,38]; as a result, lower frequency AE signals exhibit higher amplitudes.

Comparing Fig.16 and Fig.17, the spectral characteristics of the AE signals produced by the SFRC specimens subjected to 50 freeze–thaw cycles are noticeably different from those produced by specimens that were not subjected to freeze–thaw cycling. The spectral characteristics of the AE signals produced at different loading rates were relatively similar; however, the amplitudes of all frequencies decreased to some extent after the SFRC specimens were subjected to 50 freeze–thaw cycles, indicating that freeze–thaw damage causes the SFRC to lose its original physical and mechanical properties. Indeed, the concrete matrix was damaged more easily, degrading the AE signal properties during splitting tensile damage in specimens subjected to freeze–thaw cycling.

The SFRC specimens tested using the 0.02 mm/s loading rate were chosen to study the impact of fiber content on the spectral characteristics of the AE signals. Fig.18 and Fig.19 show that the changes in SFRC fiber content had minor effects on the spectral characteristics of the splitting tensile damage to the specimens both with and without freeze–thaw cycling. For all four fiber content specimens, the frequency range was 0–200 kHz and the primary frequencies in the spectra were concentrated at approximately 20, 50, and 150 kHz. The energies of the AE signals were primarily released in the 40–170 kHz range. The relative amplitude near 150 kHz tended to increase before decreasing with increasing fiber content, reaching its peak for the 1.2% steel fiber content specimen; this is consistent with the change trend of SFRC splitting tensile strength with fiber contents. A larger quantity of AE signals were assumed to be generated between the steel fibers and the concrete matrix during splitting tensile damage with a 1.2% fiber content. Comparing the spectrum images of the AE signals of specimens without and with freeze–thaw cycling, the latter appears to exhibit less activity in the proximity of the 20 kHz dominant frequency. Notably, the amplitude of the AE signal spectrum near the primary frequency of 50 kHz increased significantly owing to the generation of a weak zone within the concrete matrix—a result of the excessive incorporation of steel fibers, which led to large-scale microcracking during loading.

4 Conclusions

A series of dynamic splitting tensile tests was conducted on SFRC Brazilian disc specimens subjected to 0 and 50 freeze–thaw cycles. The effects of the loading rate and steel fiber content on the dynamic splitting tensile mechanical and damage properties of the specimens were investigated using AE techniques. The following conclusions were obtained.

(1) As the fiber content or loading rate increased, the dynamic splitting tensile strength of SFRC generally increased, but the rate of its increase eventually slowed to exhibit even a slight decrease. Furthermore, the strength of SFRC decreased after freeze–thaw cycling. The freeze–thaw damage factor decreased as the loading rate increased.

(2) The high-energy AE signals generated in plain concrete were primarily concentrated in the late stage of loading, whereas the high-energy signals generated in SFRC were distributed throughout the entire loading process. The decline in the energy counts of the AE signals produced by plain concrete after freeze–thaw cycling was more apparent than in those produced by SFRC. The proportion of low-frequency AE signals produced by SFRC during the splitting tensile tests decreased as the loading rate increased and after 50 freeze–thaw cycles, whereas the proportion of high-frequency signals increased.

(3) The SFRC generated more tensile cracks in the early stage of dynamic splitting tensile testing and exhibited more shear cracks in the later loading stage when not subjected to freeze–thaw cycling. The proportion of shear cracks in the later stage increased with increasing fiber content. Thus, freeze–thaw action had little effect on the types of cracks generated in SFRC during the early and middle stages of loading. However, more tensile cracks were generated during the later stage of loading for specimens subjected to freeze–thaw cycling.

(4) The amplitudes of all AE signal frequencies produced during the splitting tensile tests decreased in SFRC specimens subjected to freeze–thaw cycling compared with those in the specimens not subjected to freeze–thaw cycling. The relative amplitude at a frequency of approximately 150 kHz initially increased before decreasing with increasing fiber content, and its development law was consistent with the observed variation in the splitting tensile strength of the SFRC with fiber content. Thus, freeze–thaw damage was determined to reduce the activity of signals with a 20 kHz dominant frequency in the AE spectrum.

References

[1] Wang Y, Bai Y, Xi Y P. Analytical solutions of moisture-chloride ion coupled transport in unsaturated concrete. Journal of the American Ceramic Society, 2021, 104(11): 5883–5897

[2] Liu X, Yan Z, Wang D J, Zhao Y, Niu D T, Wang Y. Corrosion cracking behavior of reinforced concrete under freeze−thaw cycles. Journal of Building Engineering, 2023, 64: 105610

[3] Wang Y Z, Yang W C, Ge Y, Liu P H, Zhang A. Analysis of freeze−thaw damage and pore structure deterioration of mortar by low-field NMR. Construction & Building Materials, 2022, 319: 126097

[4] LiL FXi Y PRailsbackR. Evaluation of Different Types of Waterproofing Membranes (Asphaltic and Non-Asphaltic) as Cost Effective Bridge Deck Barriers in Reducing Corrosive Chloride Effects. Colorado: Department of Transportation, 2018

[5] GuoG. Preparation and mechanical properties of steel fiber reinforced concrete. Journal of Functional Materials, 2020, 51(11): 1001–9731

[6] Jiang L, Niu D, Bai M. Experiment study on the frost resistance of steel fiber reinforced concrete. Advanced Materials Research, 2012, 368−373: 357–360

[7] Wang C, Wu K. Study on the mechanical properties of carbon fiber and steel fiber concrete. Journal of Building Materials, 2003, 6(3): 253–256

[8] Yong Y, Ren Q W. Experimental study on mechanical performance of steel fibre reinforced concrete. Journal of Hohai University (Natural Sciences), 2006, 34: 92–94

[9] Wang R J, Hu Z Y, Li Y, Wang K, Zhang H. Review on the deterioration and approaches to enhance the durability of concrete in the freeze−thaw environment. Construction & Building Materials, 2022, 321: 126371

[10] Nili M, Azarioon A, Danesh A, Deihimi A. Experimental study and modeling of fiber volume effects on frost resistance of fiber reinforced concrete. International Journal of Civil Engineering, 2018, 16(3): 263–272

[11] Dong F, Wang H, Yu J, Liu K, Guo Z, Duan X, Qiong X. Effect of freeze−thaw cycling on mechanical properties of polyethylene fiber and steel fiber reinforced concrete. Construction & Building Materials, 2021, 295: 123427

[12] Qiu J, Pan D, Gu S, Guan X, Zheng J, Zhang C. Study on frost resistance and damage model of steel fiber reinforced coal gangue concrete. Journal of Glaciology and Geocryology, 2018, 40: 563–569

[13] Zhou T, Xiong X, Li Y. Influence of freeze−thaw cycles on dynamic performance of steel fiber reinforced concrete. Journal of Water Resources and Water Engineering, 2021, 32(3): 167–172+178

[14] Su H Z, Tong J J, Hu J, Wen Z P. Experimental study on AE behavior of hydraulic concrete under compression. Meccanica, 2013, 48(2): 427–439

[15] Sagar R V, Prasad B K R. AE energy release during the fracture of HSC beams. Magazine of Concrete Research, 2009, 61(6): 419–435

[16] Sagar R V, Prasad B K R, Kumar S S. An experimental study on cracking evolution in concrete and cement mortar by the b-value analysis of acoustic emission technique. Cement and Concrete Research, 2012, 42(8): 1094–1104

[17] Yue J, Xia Y, Fang H. Experimental study on fracture mechanism and tension damage constitutive relationship of steel fiber reinforced concrete Technology. China Civil Engineering Journal, 2021, 54: 93–106

[18] Zhang H, Jin C, Wang L, Pan L, Liu X, Ji S. Research on dynamic splitting damage characteristics and constitutive model of basalt fiber reinforced concrete based on acoustic emission. Construction & Building Materials, 2022, 319: 126018

[19] Qiu J, Pan D, Guan X, Wang M J, Zheng J. Damage law of steel fiber reinforced coal gangue concrete based on acoustic emission technology. Bulletin of the Chinese Ceramic Societ, 2018, 37: 3338–3342

[20] Lai Y S, Xiong Y, Cheng L F. Study of characteristics of acoustic emission during entire loading tests of concrete and its application. Journal of Building Materials, 2015, 18: 380–386

[21] Topolar L, Kocab D, Pazdera L, Vymazal T. Analysis of acoustic emission signals recorded during freeze−thaw cycling of concrete. Materials, 2021, 14(5): 1230

[22] Afroughsabet V, Biolzi L, Ozbakkaloglu T. High-performance fiber-reinforced concrete: A review. Journal of Materials Science, 2016, 51(14): 6517–6551

[23] Mei M, Wu L, Wan C, Wu Z, Liu H, Yi Y. Mechanical properties of nano SiO₂ and fiber-reinforced concrete with steel fiber and high performance polypropylene fiber. Materials Research Express, 2021, 8(10): 105001

[24] Zhang C, Sun Y, Xu J, Wang B. The effect of vibration mixing on the mechanical properties of steel fiber concrete with different mix ratios. Materials, 2021, 14(13): 3669

[25] XuZ L. Concise Course of Elastic Mechanics. 4th ed. Beijing: Higher Education Press, 2013

[26] Li C, Zhu J J. Study on splitting tensile failure mechanism of concrete. Shanxi Architecture, 2015, 41: 109–111

[27] Shahidan S, Pulin R, Bunnori N M, Holford K M. Damage classification in reinforced concrete beam by acoustic emission signal analysis. Construction & Building Materials, 2013, 45: 78–86

[28] Yang J Z, Liu C W. A precise calculation of power system frequency and phasor. IEEE Transactions on Power Delivery, 2000, 15(2): 494–499

[29] Cooley J W, Tukey J W. An algorithm for the machine calculation of complex Fourier series. Mathematics of Computation, 1965, 19: 297–301

[30] de Groot P J, Wijnen P A M, Janssen R B F. Real-time frequency determination of acoustic emission for different fracture mechanisms in carbon/epoxy composites. Composites Science and Technology, 1995, 55(4): 405

[31] Gutkin R, Green C J, Vangrattanachai S, Pinho S T, Robinson P, Curtis P T. On acoustic emission for failure investigation in CFRP: Pattern recognition and peak frequency analyses. Mechanical Systems and Signal Processing, 2011, 25(4): 1393–1407

[32] Ohno K, Ohtsu M. Crack classification in concrete based on acoustic emission. Construction & Building Materials, 2010, 24(12): 2339–2346

[33] Chen Z G, Jin X Y, Jin N G. Experimental study on the acoustic emission characteristics of concrete beams during damage. Cryogenic Building Technology, 2018, 40: 37–39

[34] Gan Y X, Wu S C, Ren Y, Zhang G. Study on the evaluation index of granite splitting damage based on the rise time/amplitude of acoustic emission and average frequency value. Geotechnics, 2020, 41: 2324–2332

[35] Li Y, Liu Y, Ding C, Li Z L, Zhang X F, Ding X P. Acoustic emission RA-AF characterization of the damage process of different seam coal. Mine Safety, 2022, 53: 37–43

[36] Aggelis D G. Classification of cracking mode in concrete by acoustic emission parameters. Mechanics Research Communications, 2011, 38(3): 153–157

[37] Xing Y K, Huang B X, Chen D Y, Zhao X L, Li B H. Acoustic emission full-waveform multi-parameter monitoring of nonlinear fracture of fractured fractures. Journal of China Coal Society, 2021, 46: 3470–3487

[38] Li X L, Chen S J, Liu S M, Li Z H. Acoustic emission waveform characteristics of rock mass under uniaxial loading based on HHT analysis. Journal of Central South University, 2021, 28: 1843–1856

RIGHTS & PERMISSIONS

Higher Education Press