1. Introduction
Blasting is a common rock excavation method for water conservation, transportation, and mining industries. However, high-pressure gas generated by blasting can open and penetrate microfractures inside the rock mass, which can cause the mechanical strength of the rock mass to weaken, causing local instabilities or even damage to the rock mass. This can seriously threaten the lives of the construction crew and the normal operation of mechanical equipment, causing incalculable losses. Therefore, studying the weakening effects of a rock mass under blasting conditions is an important issue in slope blasting damage research, as well as in determining rock mass changes at the quality level.
Researchers worldwide have studied rock slope blasting damage using shaking table tests, numerical simulations, micro-seismic monitoring, and acoustic detection. For example, Jian et al. [
1] used a shaking table test based on similarity theory for an indoor jointed slope model, and the results showed that the long-term vibration load caused the microfractures to expand and the master fracture and native joints to penetrate into the jointed slope, leading to local instabilities or even damage to the overall rock mass. Li et al. [
2] used the UDEC-Voronoi method to reproduce the slope deformation and damage process based on a centrifuge shaking table test and introduced an entropy value to quantitatively analyze slope damage and damage characteristics. Fei and Yuan [
3] used ANSYS to perform numerical simulations to study the change law of the safety coefficient during blasting, based on the strength reduction and time-course analysis methods. Liu [
4] used LS-DYNA to perform three-dimensional finite element calculations to assess the blasting damage on high slopes containing structural surface rocks. For micro-seismic monitoring, Gong et al. [
5] established an on-site micro-seismic monitoring system to detect the degree of damage caused by blasting vibrations to the internal rock slopes; this study verified the feasibility of using micro-seismic technology for monitoring the stability of rocky slopes. A study by Zhang et al. [
6] also showed that apparent micro-seismic stress events could reflect the degree of blasting damage during construction. In addition, Yan et al. [
7] established a connection between acoustic wave velocity and the mechanical properties of the rock, proving that variations in acoustic wave velocity could effectively characterize the degree of rock damage. Wu [
8] also used acoustic wave detection to study the blasting damage of the surrounding rocks in the Badaling tunnel blasting. Song et al. [
9] proposed a dynamic stress ratio evaluation method that more comprehensively reflected the damage degree of the geotechnical structures subjected to blasting vibration. Chu et al. [
10] conducted cumulative damage and blast vibration tests on concrete specimens to further determine the blast vibration wave propagation and attenuation based on cumulative damage. Another study by Zhang et al. [
11] combined vibration and fractal theories to explore the evolution of tensile rock damage and rupture mechanisms. Sun et al. [
12] studied rock fracture damage and the bench blast vibration law during blasting using a continuous damage intrinsic model. Zhang et al. [
13] introduced cumulative damage of rock blasting and established a multivariate nonlinear mathematical model for the blasting vibration attenuation law under cumulative rock mass damage. Liu et al. [
14] showed that the concrete double K criterion could also be used to study the cumulative damage effect on slopes. However, all of the above-mentioned studies did not address the cumulative nature of damage that occurs as a result of blasting rocky slopes, and few studies have been published on attenuating the mechanical properties of slopes as a result of damage.
To quantitatively investigate the effect of blasting on the mechanical properties of rock masses, Hoek et al. [
15] introduced a disturbance factor
D in the H–B strength criterion, reflecting the changes in the mechanical properties of the rock after blasting. Furthermore, researchers have attempted to determine the disturbance factor values
D,
mb, and
s in the H–B strength criterion. Yan et al. [
16] introduced the rock integrity factor
KV and blasting damage factor
D in the Hoek–Brown criterion to establish how
mb and
s can be used to characterize the cumulative damage effects of rock blasting, as well as determine the state of rock blasting disturbance and the degree of damage to the mechanical properties of the rock. Shi et al. [
17] combined the disturbance factor and damage degree based on the fuzzy theory, and used the acoustic wave velocity of the rock to establish a link between the sound velocity reduction rate, disturbance factor, and the damage degree to optimize and improve the H–B criterion. Xia et al. [
18] established an equation for estimating GSI and the disturbance factor
D from the rock wave velocity. Li and Xue [
19] proposed a method for solving the disturbance factor
D and geological strength index GSI by using the wave velocities of disturbed rock and rock longitudinal wave velocities, combining them into the Hoek–Brown criterion to obtain a new equation. Results from Haghnejad et al. [
20] also showed that the blasting damage factors
D and GSI play important roles in slope damage. Yang et al. [
21] used the rock acoustic P-wave velocity as a supplemental element to estimate the disturbance factor
D and compensate for the disturbance factor
D in the H–B failure criterion. Yang et al. [
22] investigated the relationship between the disturbance factor
D and the damage depth based on rock acoustic wave velocity test results after blasting and studied the detrimental effects on the mechanical properties of the rock.
Considering that the root cause of rock damage due to blasting excavation is a reduction in the mechanical properties of the rock, acoustic wave velocity changes can effectively be used to characterize rock damage. While research has been conducted on the cumulative effects of blasting excavation on reserved rock, few studies have focused on the reduction in mechanical properties due to blasting excavation. Furthermore, most studies have not established a direct link to the degree of rock damage but have characterized the rock damage indirectly through rock wave velocity values. Therefore, a sufficient theoretical understanding is lacking, specifically on the correlation between rock damage theory and rock mechanics, and this is not conducive to the development of the Hoek–Brown criterion or estimating the mechanical properties of disturbed rock masses.
In this work, we derived expressions for the damage degree Ds and disturbance factor D in rock damage theory, by combining the Hoek–Brown criterion with a previously established numerical model based on the RHT damage principal structure, focusing on the decay law of the rock mechanical properties under different damage conditions. The results of this study could provide a theoretical basis for the protection of rocky slopes under blasting conditions and establish stability evaluations.
4. Results and Discussion
4.1. Analysis of the Blast Damage Results
Figure 3 and
Figure 4 show the schematic of the calculated results from the blasting excavation damage model and the schematic of the bench part damage, respectively.
Damage to the reserved rock mass was mainly caused by adjacent rock blasting excavation. Eventually the adjacent reserved rock mass bench slope was damaged, caused by a larger damaged area. With increasing distance between the burst center, the degree of damage to the reserved rock mass gradually declined. During explosive detonation, the blast generated by the shock wave blast load instantly reach a peak and the shell hole wall around the rock was instantly crushed, forming a crushing circle. The section near the shell hole wall portion of the rock contained a greater degree of damage, and the damage degree value was close to 1. With an increase in the burst core distance, the shock wave gradually decayed into a stress wave. The stress wave did not directly damage the rock, but it caused new cracks in the rock mass and primary cracks continued to extend and expand with penetration, causing varying degrees of damage to the rock. Therefore, the damage degree values ranged between 0.2 and 0.8. When the stress wave further decayed, and the circumferential tensile stress in the rock was less than the tensile strength of the rock, rock crack development gradually stopped. Therefore, this formed the undamaged rock mass area.
To obtain
Ds = 0.592, we substituted GSI = 55 and
D = 0 into Equation (9). When the rock GSI had a value other than 100, a disturbance factor of
D = 0, and the rock damage degree was not equal to 0, initial damage to the rock mass, caused by fractures in the native rock mass before blasting excavation, was assumed. Similarly, to obtain
Ds = 0.91, we substituted GSI = 55 and
D = 1 into Equation (9). Thus, when the perturbation factor was 1, the maximum rock damage value was 0.91 and not 1. As shown in
Figure 5, the damage degree
Ds range (0.592, 0.910) with increasing perturbation factor
D exhibited an increasing logarithmic function for the damage degree
Ds, and the growth rate slowed.
4.2. Determining Rock Mechanical Properties after Blasting Excavation
From the calculated results, we randomly selected 12 sets of data; their damage degrees were 0.616, 0.632, 0.672, 0.699, 0.722, 0.742, 0.782, 0.803, 0.832, 0.867, 0.892, and 0.910. The corresponding disturbance factor
D values were obtained by using a calculation program within MATLAB R2020b (Yunnan, China). The
D values were input into Equations (2)–(10) to obtain the mechanical properties of the rock mass under different disturbance conditions (i.e., under different damage degrees); the calculated results are shown in
Table 5 and
Table 6.
The following rock mass mechanical property ratios
,
,
Ermi/
Erm0,
crmi/
crm0,
φrmi/
φrm0 were selected to evaluate the attenuation degrees of the mechanical properties for the blasted and undisturbed rock bodies. Smaller ratios indicate greater degrees of property attenuation, as shown in
Figure 6.
The tensile strength of a rock mass declines almost linearly with increasing damage, and the compressive strength, modulus of elasticity, cohesion, and internal friction angle values decreased exponentially with increasing damage. In terms of mechanical property decline, the angle of internal friction decreased slowly with increasing damage, followed by the cohesion and modulus of elasticity. Compressive strength decreased the most rapidly with increasing damage.
4.3. Evolution of Blasting Excavation Slope Damage
The time course diagram for blasting excavation slope damage is shown in
Figure 7. The blasting damage first occurred at the bottom of the reserved rock and the top of the adjacent benches; then, the damage gradually penetrated the bottom. As the blasting process proceeded, a distinct damage zone appeared at the slope bench and gradually extended to the interior of the bench until blasting was complete. Eventually, a damage zone of 1.5 to 2 m in depth formed at the adjacent bench and the bottom portions of the reserved and stripped rocks.
When the numerical calculation result of the rock damage degree is 0 < Ds < 0.318, the actual damage degree of the rock is 0.592–0.910; when the numerical calculation result is greater than or equal to 0.381, the actual damage degree of the rock is the maximum value of 0.910; that is, the slope has been completely damaged.
4.4. Rock Damage Classification Criteria
Based on Equation (14), the rock level was classified into four classes: extremely stable (I), stable (II), moderately stable (III), and unstable (IV), as shown in
Table 7.
By combining Equations (11) and (13), we can obtain the basic quality index of the rock mass BQ, according to the correction method for rock mass level slope engineering, following the standard for engineering rock mass classification [
24]:
where
λ is the correction coefficient of the main structural surface type and extension extensibility for slope engineering,
K4 is the correction coefficient for groundwater influence for slope engineering,
K5 is the correction coefficient for the main structural surface yield effect for slope engineering,
F1 reflects the influence of the relationship between the inclination of the main structural surface and the inclination of the slope,
F2 reflects the influence of the main structural surface,
F3 reflects the relationship between the request of the slope and the inclination of the main structural surface, and the coefficient
F3 reflects the influence of the relationship between the slope request and the inclination of the main surface structure. After statistical analysis,
λ was 0.7,
K4 was 0.2,
F1 was 0.4,
F2 was 1.0, and
F3 was 0.2. The
Ds values selected were 0.592, 0.672, 0.832, and 0.91. The calculated results are shown in
Table 8.
Lithology is a function of harder rock as denoted by II-V, which can be used to grade the basic quality of the rock mass, corresponding to I-IV for the self-stabilizing ability grading of the rock mass [
24]. Therefore, the damage degree
Ds was used as the standard to determine the self-stabilizing grading of the rock mass.
The initial damage degree of the slope was Ds = 0.592, which was medium-stable. However, under blasting action, a damage zone with a depth of ~1.5–2 m formed on the slope surface, and the damage degree of the rock in the damage zone was Ds > 0.6, which corresponded to rock mass grades III and IV.
After blasting excavation, the damage degree Ds > 0.82 corresponded to rock mass grade IV. After the rock mass broke, we observed broken belts, falling blocks, and dumping, which pose significant safety risks to personnel and equipment. When the damage degree Ds was (0.6, 0.82), it corresponded to a rock mass grade of III, where the rock mass was broken, exposing the rock bodies on the slope bench surfaces for that range, with some surfaces occasionally containing falling blocks. Therefore, if necessary, reinforcement measures such as slurry spraying and anchor net hanging should be implemented on these slope surfaces. As the damaged area caused by blasting deepened, the actual width of the slope bench was less than the design width, and over-excavation occurred; therefore, additional actions should be taken such as reducing the size of the hole and changing the loading structure.
5. Conclusions
By introducing a link between the damage degree and the disturbance factor in the Hoek–Brown strength criterion, we quantitatively determined a law for the degree of rock mass damage as a function of disturbance factor changes. We established a numerical model based on data from a phosphate mine slope in Guizhou Province by applying the RHT damage principal model. We then assessed the changes in mechanical properties of the rock under different damage conditions and the following conclusions were drawn.
The tensile strength of the rock mass decreased linearly with increasing damage degree Ds, and the compressive strength, elastic modulus, cohesion, and internal friction angle values of the rock mass decreased non-linearly with increasing damage degree. The internal friction angle exhibited the smallest decline, followed by cohesion and elastic modulus, while compressive strength decreased the most rapidly.
Based on the standard for engineering rock classification and calculations for the damage degree, we established a rock mass self-stabilizing classification standard based on the damage degree Ds. The rock mass self-stability of the mine was classified into four grades: extremely stable (I), stable (II), moderately stable (III), and unstable (IV).
Considering the initial damage to the rock mass, the damage state of the surrounding rock should be considered in the blasting design, and the amount of charge should be appropriately reduced. An appropriate charging structure should also be designed to reduce the damage to the reserved rock mass. Furthermore, after blasting excavation, the slope surface should be properly prepared, and if necessary, the reserved rock mass should be reinforced by slurry spraying or anchor hanging.
In this work, we established a link between damage degree and disturbance factor, and the mechanical properties of the rock mass were derived from the disturbance factor. However, there were limitations in the numerical simulation when we obtained the exact damage depth of the rock mass. Therefore, in future research, more field tests should be conducted and combined with the research results from this study to establish damage degree, damage depth, disturbance factor, and mechanical index values. We also plan to conduct numerous field tests to establish a link between damage degree, damage depth, disturbance factor, and the mechanical index, in order to provide better theoretical support for the estimation of rock mechanics parameters and rock protection in the future.