Abstract

Sudden inelastic deformations in rock are associated with acoustic emission (AE). Therefore, AE monitoring technique can be used to study the fracture processes of rock. In this paper, AE tests were conducted on the granitic gneiss specimens under the uniaxial compressive loading conditions. The temporal changes in AE hit parameters and spatial-temporal evolution of AE events during the failure process of the granitic gneiss specimens were studied, and several characteristic AE phenomena (i.e., dramatic increase in dominant frequency, AE energy, and hit rate, the AE event with a high energy level, and the through-going distribution of the AE events with intermediate energy levels) were statistically analyzed before the failure occurred. It was found that the chronological order of the characteristic AE phenomena was relatively certain and correspondingly had a close relationship with the crack development stage. Because of the difference of the stress level at each crack development stage, the stability at different crack development stages is different. Therefore, a rock stability assessment approach was established based on the chronological order of the characteristic AE phenomena, and then the rock stability was assessed using the proposed approach.

1. Introduction

As a brittle material, rock will experience sudden inelastic deformations before failure, such as the initiation and propagation of cracks when it is subjected to external loads. These deformations are associated with acoustic emission (AE) which is defined as a transient elastic wave generated by the rapid release of strain energy [1]. Therefore, AE monitoring technique can be used to study the fracture or failure process of rock.

Since the 1960s, considerable efforts [26] have been made to analyze the evolution characteristic of AE and recognize the characteristic AE phenomena to assess the rock stability and forecast its failure.

The number of AE hits is most widely analyzed among all the AE parameters because the AE hit is easy to acquire, and its value has a positive correlation with the crack number. Many previous studies have shown that, when approaching the peak stress, the AE hit rate increases significantly in the rock whose plastic deformation is not obvious, while the AE hit rate might decrease in the rock whose plastic deformation is obvious [7]. These results indicate that the AE hit rate is significantly influenced by the rock deformation characteristic.

The AE energy is another AE parameter that is widely analyzed; the significant increase of AE energy before rock failure is observed in a large number of rock types, such as granite [8, 9], gneiss [10], and tuff [11], indicating that the increase of energy is a reliable precursor of the rock failure.

In addition, research studies on the spectral analysis of AE have been conducted in the past decades. For example, Spetzler et al. [12] found that, as failure was approached, more power was recorded at the lower frequency in large granite and basalt specimens. He et al. [13] observed that there were much lower frequency AE events near the bursting failure of the limestone specimens. However, Li et al. [14] observed the increase of the higher frequency component before rock failure under cyclic loading and multistage loading. Ji et al. [15] have also observed the similar phenomena in granite and marble. The conflicting experimental results indicate the relationship between the spectrum and fracture in rock is too complex to be well summarized.

Further, detailed analysis of the space-time distribution of AE events can help us better understand the rock fracture process [1]. For example, if the failure was induced by through-going shear fault or compaction band, then, before the failure occurred, the AE events tended to gather around the future macroscopic fracture plane [1618]. Thus, the spatial distribution of AE events can be used for not only failure warning but also macroscopic crack shape forecasting. Nevertheless, in some relatively homogeneous rock, AE events exhibited dispersion when failure occurred, and the macroscopic crack shape cannot be forecasted based on the spatial distribution of AE events before the failure occurred [19].

Based on the above introduction, it can be found that, many characteristic AE phenomena have been recognized. However, the temporal relation between each characteristic AE phenomenon has rarely been reported in the literature. Therefore, in the present study, AE tests were conducted on granitic gneiss specimens under uniaxial compressive loading conditions, and then the chronological order (temporal relationship) of characteristic AE phenomena was studied. On this basis, a rock stability assessment approach was established.

2. AE Tests of Granitic Gneiss Specimens

2.1. Experimental System

The experimental system consists of the load device and the monitoring apparatus. The load device used the computer control electrohydraulic servo press TAW-2000KN. PCI-2 AE test apparatus was applied to collect the AE signals produced by granitic gneiss. In this experiment, eight Nano30 sensors were arranged on the specimen surface (the small grey cylinders). 45 dB threshold was selected for all sensors, and the preamplifier gain was 40 dB. Each waveform was digitized into 1024 samples at a sampling rate of 1 MHz.

2.2. Rock Specimen and Loading Control

The granitic gneiss specimens (Φ50 mm × 100 mm) used in the AE experiment were collected from the Shirengou iron mine, Tangshan city, Hebei province, China. The relatively detailed introduction about the mine can be found in the literature [20]. The granitic gneiss consists of approximately 40% plagioclase, 32% alkali feldspar, 8% quartz, 5% biotite, and 15% hornblende [21]. Uniaxial loading was conducted on 12 specimens (No. Sim1∼12), and the loading rate was 0.003 mm/s.

3. Characteristic Parameters of AE

3.1. AE Hit Parameters

A typical AE hit is shown in Figure 1. The “threshold” is a preset voltage value; i.e., only the signal whose voltage value is higher than the threshold can be detected by the AE sensor. The “hit” is a detected signal. The “hit rate” is the number of hits per second. The “energy” is based on the sum of the squared voltage readings divided by a token resistance R (R is equal to 10 kΩ). The energy is reported in attojoules (aJ, 1 aJ = 10−18 J).

Using the fast Fourier transformation (FFT), the AE hit can be converted from the time domain to the frequency domain, and the dominant frequency is the corresponding frequency of the maximum voltage (Figure 1).

3.2. AE Event Parameters

One of the major strengths of AE technique is its ability to locate an active source in three dimensions when enough hits are available from multisensors. The located source is called an AE event in this paper.

The energy of each AE event is determined as follows:where is the absolute energy of the AE hit detected by each sensor, is the distance between the AE event and the sensor in millimeters, and is the total number of sensors used for the energy calculation. The computed value is an average energy for the whole array, assuming elastically propagating, spherical waves of a point source, corrected for geometrical spreading on a 10 mm reference sphere.

The energy level of each AE event is determined as follows:

4. The Change of AE Hit Parameters: Experimental Results

To seek the change characteristic of AE during the fracture process of rock specimens, the temporal changes in parameters of AE hits received by a single sensor are studied.

According to the difference in the change characteristic of the AE hit parameters, the specimens can be divided into two categories. One category includes the majority specimens and can be represented by specimen Sim2. Another category includes a minority of the specimens and can be represented by specimen Sim10. In this section, the specimens Sim2 and Sim10 were taken as examples to illustrate the change characteristic of the AE hit parameters of the granitic gneiss specimens. It should be noted that, the change characteristics of the AE hit parameters of the other specimens were similar with specimen Sim2 or Sim10 besides the ranges of the AE hit parameter values being somewhat different. To avoid the redundancy of illustration, the change characteristics of the AE hit parameters of the specimens have not been discussed one by one.

4.1. Hit Rate

As shown in Figure 2(a), the specimen Sim2 experienced the crack closure stage (I, 0%–17% peak strength), the elastic stage (II, 17%–51% peak strength), the stable crack growth stage (III, 51%–77% peak strength), and the unstable crack growth stage (IV, 77%–100% peak strength), successively. The hit rate buildup commenced at 230 s–240 s, which corresponded to the transition from the elastic stage to the stable crack growth stage, and became noticeable at approximately 300 s, when the rock entered the unstable crack growth stage.

For the specimen Sim10 (Figure 2(b)), a different trend in the AE hit record was observed; a noticeable increase of the hit rate not only occurred at a high stress level (such as the unstable crack growth stage) but also occurred at a low stress level of only 6%–15% peak strength. We interpret this special AE response at low stress level as unstable crack growth in a local region of the specimen, which will be confirmed by the spatial distribution of AE events in Section 5.3. Therefore, the special AE response stage was named “local unstable crack growth stage (L)” in this paper. Furthermore, it can be found that the maximum hit rate might be reached at the local unstable crack growth stage or the unstable crack growth stage.

4.2. Dominant Frequency

The changes of dominant frequency shown in Figure 3 were calculated using a moving window approach. For both the calculation window and the sliding window, 100 hits were adopted. The arrival time of the last hit in the calculation window was used as the time of the calculated result.

As shown in Figure 3, for both Sim2 and Sim10, the dominant frequency increased significantly with stress in the low stress level. Before specimen failure occurred, different change trends of the dominant frequency, such as decrease (Figure 3(a)) and fluctuation (Figure 3(b)), were observed in different specimens. In addition, no obvious correlation was observed between the change trend before the failure and whether the specimen experienced the local unstable crack growth stage.

4.3. Energy

The changes of energy were calculated using the same approach in Section 4.2. As shown in Figure 4(a), the energy of specimen Sim2 fluctuated in a low level at the crack closure stage (I), the elastic stage (II), and the stable crack growth stage (III). In contrast, a drastic increase of energy occurred at the unstable crack growth stage (IV). For the specimen Sim10 (Figure 4(b)), the energy increased dramatically at the local unstable crack growth (L) and the unstable crack growth stage (IV).

5. The Spatial-Temporal Evolution of the AE Events: Experimental Results

According to the difference in the spatial-temporal evolution of the AE events, the specimens can be divided into three categories. In this section, the specimens Sim2, Sim9, and Sim10 were taken as examples to illustrate the three categories.

5.1. The General Spatial-Temporal Evolution of the AE Events

The evolution characteristic of the AE events in a majority of specimens is similar to that of specimen Sim2. As shown in Figure 5, the spheres represent the AE events and the different colors indicate the different energy levels. The top 20% of the energy level scale is defined as the high energy level (red and orange), the bottom 45% of the energy level scale is defined as the low energy level (blue and purple), and the intervening 35% is defined as the intermediate energy level (yellow and green).

According to the location results (Figure 5), the AE events trend from the top and bottom to the middle of the specimen. Before 318 s, AE events with low energy level and intermediate energy level were concentrated in the top and bottom of the samples, and there were no AE events located in the middle of the sample (Figure 5(a)). At 323.7 s, the first AE event with a high energy level occurred (Figures 5(b) and 6), corresponding to approximately 85% of the peak strength (Figure 2(a)). From 318 s, the number of AE events with an intermediate energy level began to increase in the middle of the specimen and formed a through-going distribution (Figures 5(b) and 5(c)). At 326 s, an AE event with a high energy level occurred in the middle of the specimen (Figures 5(c) and 7). After approximately 18 s, specimen failure occurred (Figure 5(d)).

From the general spatial-temporal evolution of the AE events, which can be represented by Sim2, it can be inferred that the three phenomena, the first AE event with the high energy level, the through-going distribution of AE events with the intermediate energy level, and the AE event with the high energy level in the area which was the last through by AE events with the intermediate energy level, occurred in sequence at the high stress level.

The phenomenon that the number of AE events in the middle of the sample was still few at the high stress level (Figure 5(b)), might be mostly attributed to the loading boundary effects such as friction and uneven stress on the specimen ends. The friction was due to elastic parameter mismatch between the loading platen and the specimen, and the uneven stress was due to the unflatness of the specimen ends. This observation is consistent with experimental observations made in [17, 22, 23]. Stress concentration at the specimen ends is believed to be the consequences of the fractures generated at the top and bottom of the specimen. And as the load increases, the fractures propagated from the top and bottom to the middle of the specimen.

In addition, the phenomenon that the spatial distribution of AE events exhibited dispersion before the failure occurred was mainly due to the relatively homogeneous strength distribution in the specimens. The main mineral components of granitic gneiss specimens used in this paper are plagioclase, alkali feldspar, and hornblende, which account for more than 85%. The hardness of plagioclase, alkali feldspar, and hornblende is about 6, and the minerals with similar hardness are generally similar in strength. Therefore, granitic gneiss specimens are relatively homogeneous in spatial distribution of strength. As the experimental results shown in the literature [19, 24, 25], in some relatively homogeneous rock, the spatial distribution of AE events exhibited dispersion and cannot forecast the macroscopic crack shape before the failure occurred.

5.2. The Gap of AE Event with Intermediate Energy Level

The evolution characteristic of AE events at the high stress level in some specimens is different from the general evolution characteristic, as illustrated by specimen Sim9.

The spatial-temporal distribution of AE events in Sim9 is shown in Figure 8. At 344 s, two AE events with a high energy level occurred in succession at the top of the specimen (Figures 8(a) and 9), corresponding to approximately 88% peak strength. The phenomenon, i.e., the AE events with a high energy level occurred at a high stress level, in agreement with the general evolution characteristic is described in Section 5.1. However, when the failure occurred at 364 s, there was still a gap of the AE event with an intermediate energy level, and in the gap, the AE events with low energy level were few (outlined by dashed lines in Figure 8(c)). This gap suggested that the microfractures in the gap were few when approaching failure and that the macroscopic crack (Figure 8(c)) passed through the gap suddenly with an obvious brittle failure characteristic when failure occurred.

5.3. The AE Events with a High Energy Level at a Low Stress Level

The evolution characteristic of AE events at low stress levels in some specimens is different from the general evolution characteristic. The specific evolution characteristic only appears at the local unstable crack growth stage, as illustrated by specimen Sim10.

The spatial-temporal distribution of AE events in Sim10 is shown in Figure 10. The local unstable crack growth stage lasted from approximately 128 s to 231 s, corresponding to Figures 10(b) and 10(c). At 163 s, the first AE event with a high energy level occurred at approximately 8.2% peak strength (Figures 10(b) and 11), which was much lower than the majority specimens. In the meantime, the AE events gathered in a band shape around the AE event with a high energy level, suggesting there was a weak area where the strength was significantly lower than that in the other area and the fracture of the weak area induced the high energy level AE event. In addition, as the load applied on the specimen was still low during the local unstable crack growth stage, the density and energy level of AE events were low in the area far away from the weak area.

6. Chronological Order of the Characteristic AE Phenomena

Based on the above analysis, several characteristic AE phenomena during the failure process of the specimens can be observed, such as the obvious increase in the hit rate, dominant frequency, and energy. The chronological order of these AE phenomena and the corresponding crack development stage are summarized in Table 1.

As shown in Table 1, the obvious increase in dominant frequency always occurred first among all the characteristic AE phenomena, and it occurred before the specimens entered the unstable crack growth stage.

The obvious increase in the energy and the hit rate and the first AE event with a high energy level (short for “first high” in Table 1) tended to occur simultaneously or at short intervals. In consideration of the observation that the energy and hit rate have positive correlations with the microfracture scale and crack number, respectively, and it can be inferred that the crack area and crack number tend to increase simultaneously or at short intervals.

For most of the specimens, the first AE event with a high energy level might occur either at the local unstable crack growth stage or at the unstable crack growth stage. If the first AE event with a high energy level occurred in the unstable crack growth stage, there are two types of AE event distribution characteristics: (1) AE events were few and scattered, such as for Sim8 (Figure 12), which indicated that the rock was relatively uniform and there were few new microfractures formed under the high stress level and (2) many AE events with intermediate energy levels occurred, which indicated that the rock was relatively nonuniform and there were many new microfractures formed under the high stress level, such as for Sim2 (Figures 5(b) and 6) and for Sim9 (Figures 8(a) and 9).

If the first high energy level AE occurred in the local unstable crack growth stage, it can be found that some AE events with intermediate energy levels gathered around the AE event with a high energy level, and the density and energy level of the AE events were low in the area far away from the AE event with the high energy level (Figure 10(b)). Thus, the spatial distribution characteristic of the AE events can help distinguish between the local unstable crack growth stage and the unstable crack growth stage.

The through-going distribution of AE events (short for “through-going distribution” in Table 1) and the AE event with a high energy level in the area that was the last to undergo AE events with intermediate energy levels (short for “high in last to undergo” in Table 1) would occur at the unstable crack growth stage (IV) and after all the characteristic changes of the AE hit parameters. For some specimens, such as Sim8, Sim9 (Figure 8(b)), and Sim12, the failure occurred before the through-going distribution of the AE events with intermediate energy levels. However, if the through-going distribution of the AE events with intermediate energy levels occurred, then an AE event with a high energy level would be certain to occur in the area that was the last to undergo AE events with intermediate energy levels.

7. Rock Stability Evaluation Based on the Chronological Order of the Characteristic AE Phenomena

7.1. Rock Stability Evaluation Approach

Based on the statistical result (Table 1) and analysis above, the general chronological order of the characteristic AE phenomena and the corresponding crack development stage are shown in Figure 13. Because of the difference of the stress level at each crack development stage, the stability at different crack development stages is different. Thus, based on the chronological order of the characteristic AE phenomena, the rock stability can be assessed.

The rock stability evaluation approach is as follows:(1)The obvious increase in dominant frequency indicates that the specimen has not entered the unstable crack growth stage, i.e., the stability is high to medium.(2)When the obvious increase in energy and hit rate and the first AE event with a high energy level occurred, if the AE events gathered in a band shape around the AE event with a high energy level, and the density and energy level of AE events were low in the area far away from the AE event with a high energy level, then the rock has entered the local unstable crack growth stage. Because the stress level is still low during the stage, the stability can be assessed as high.(3)When the obvious increase in energy and hit rate and the first AE event with a high energy level occurred, if there were rare AE events or many AE events with intermediate energy levels, then the specimen has entered the stable crack growth stage or the unstable crack growth stage, indicating that the rock stability is low.(4)AE events with intermediate energy level forming the through-going distribution indicate that the microfractures have fully developed and the rock stability is very low.(5)The occurrence of the AE event with high energy level in the area that was the last through by AE events with intermediate energy levels is the last characteristic AE phenomenon before the rock failure, i.e., failure is approaching.

7.2. Rock Stability Evaluation Result

Taking Sim2 as an example, the rock stability evaluation can be illustrated. As shown in Figure 14, the dominant frequency began to increase at 30 s, and the stability was assessed as high. From 300.1 s to 324.2 s, obvious increases were observed in energy and hit rate, and the first AE event with high energy level occurred successively, i.e., the stability was low. From 324.3 s to 325.9 s, the number of AE events with intermediate energy level increased in the middle of the specimen and formed a through-going distribution (Figure 14(c)), suggesting that the rock stability was very low. At 326 s, an AE event with high level occurred in the middle of the specimen (Figure 14(d)), i.e., failure was approaching.

8. Conclusions

AE monitoring was used to study the fracture process of granitic gneiss under the uniaxial loading condition. A rock stability assessment approach was established based on the chronological order of the characteristic AE phenomena. The following conclusions can be drawn:(1)Under the uniaxial loading condition, several characteristic AE phenomena, such as dramatic increases in dominant frequency, energy, and hit rate, the AE event with a high energy level, and the through-going distribution of AE events with intermediate energy levels, were observed before failure occurred, thus indicating that AE monitoring has the potential to assess the rock stability.(2)For most of the granitic gneiss specimens, the chronological order of the characteristic AE phenomena was certain and had a corresponding relationship with the crack development stage. Because of the different stress levels at each crack development stage, the stability at different crack development stages is different. Thus, based on the chronological order of the characteristic AE phenomena, the rock stability can be assessed.

Data Availability

The AE monitoring data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

The work was financially supported by the National Key Research Project (2016YFC0801607), the National Natural Science Foundation of China (51604062 and 51574060), and the Science and Technology Major Project of Anhui Province (17030901023). The authors are thankful to the referees and editors for their valuable comments and suggestions devoted to improving the quality of our manuscript.