Experimental investigation on dynamic strength and energy dissipation characteristics of gas outburst‐prone coal

We report laboratory experiments to investigate the dynamic failure characteristics of outburst‐prone coal using a split Hopkinson pressure bar (SHPB). For comparison, two groups of experiments are completed on contrasting coals—the first outburst‐prone and the second outburst‐resistant. The dynamic mechanical properties, failure processes, and energy dissipation of both outburst‐prone and outburst‐resistant coals are comparatively analyzed according to the obtained dynamic compressive and tensile stress‐strain curves. Results show that the dynamic stress‐strain response of both outburst‐prone and outburst‐resistant coal specimens comprises stages of compression, linear elastic deformation, then microfracture evolution, followed by unstable fracture propagation culminating in rapid unloading. The mechanical properties of both outburst‐prone and outburst‐resistant coal specimens exhibit similar features: The uniaxial compressive strength and indirect tensile strength increase linearly with the applied strain rate, and the peak strain increases nonlinearly with the strain rate, whereas the elastic modulus does not exhibit any clear strain rate dependency. Differences in the dynamic failure characteristics between outburst‐prone and outburst‐resistant coals also exist. The hardening effect of strain rate on outburst‐prone coal is more apparent than on outburst‐resistant coal, which is reflected in the dynamic increase factor at the same strain rate. However, the dynamic strength of outburst‐prone coals is still lower than that of outburst‐resistant coals due to its low quasi‐static strength. The dissipated energy of outburst‐prone coal is smaller than that of outburst‐resistant coal. Therefore, the outburst‐prone coal, characterized by low strength, high deformability, and small energy dissipation when dynamically loaded to failure, is more favorably disposed to the triggering and propagation of gas outbursts.


| INTRODUCTION
Coal is the basic energy and fuel source in China. Coal and gas outbursts during underground coal mining are characterized by the sudden ejection of large masses of coals or volumes of gases. [1][2][3][4] Such outbursts may result in significant damage to equipment ( Figure 1A) and may also result in injury and fatalities. [5][6][7][8] Since the first-recorded outburst happened in the Isaac coal mine in Lule coal field in 1843, more than 40 000 outbursts have been reported worldwide. 9 The most disastrous mine outbursts resulted in 187 deaths in the Nowa Ruda Colliery (Poland) in 1941, 148 deaths in the Daping coal mine (China) in 2004, and 214 deaths in the Sunjiawan coal mine (China) in 2005. 2,10,11 Much exploratory research over the past 100 years has probed this catastrophic phenomenon to obtain an improved understanding of the causative mechanisms and better mitigate outbursts. 6,[12][13][14][15][16][17][18] However, few of the proposed mechanisms can be successfully applied to explain, predict, and prevent outbursts. Due to the unique characteristics of these destructive, short duration, and high-intensity events, outbursts are both unpredictable in space and time. 5,6,19,20 It remains difficult to accurately measure in situ stress and gas pressure or to examine coal properties at different scales and for varied impact loads. 21 Current wisdom is that outbursts are a violent dynamic instability of the coal mass when the interaction of abnormal geostresses, high gas pressures, and the presence of "outburst-prone" coal reaches a critical condition. 22 The divergence of different theories lies in the myriad of different reasons that this critical condition is reached. These include the stress-concentrating influence of faults, folds, the presence of gas pockets, the impacts of thickness variation in the coal seam, among others. 23,24 Although these theories commonly link the propensity for outbursting to the presence and action of "outburst-prone coal" (Figure 1), the key features of its dynamic failure characteristics under impact loading remain poorly understood and constrained. [25][26][27][28][29][30][31][32] Although the quasi-static physical properties of coal and rock have been studied widely, the dynamic failure characteristics of coal and rock are less fully constrained. 5 Dynamic properties of interest include the following: density, wave velocity, porosity, strength, scale effect, bedding effect, the influence of moisture, and energy dissipation. 15,[33][34][35][36][37][38][39][40][41][42][43][44][45] Very high loading rate dynamic tests are usually conducted by using split Hopkinson pressure bar (SHPB) systems, 30,46 to determine dynamic properties of brittle materials including concretes, ceramics, rocks, and coals under wide range of impact loadings or strain rates of 10 1 -10 4 /s. [47][48][49][50][51][52][53] For determination of dynamic properties of different coals, the principal attributes are to recovery the dynamic magnitudes of Young's modulus and compressive/tensile strength based on the stress-strain curves at very high strain rates. 43,45 For instance, scholars have investigated the stress-strain curves and energy dissipation of coals under different bedding orientations, dry or saturated coals. [54][55][56] Fractal characteristics and electromagnetic radiation of crack propagation of coal were tested during an SHPB test. 57,58 So far, extensive suites of dynamic SHPB tests have been carried on rock and common coal materials, but only a few such dynamic tests have been completed on outburst-prone coals. 59 The diversity of dynamic characteristics between prone and resistant coals was not reported. Here, we supplement this dearth of observations by recovering a full suite of dynamic failure and energy dissipation characteristics to contrast the response of outburst-prone coals against a control sample of outburst-resistant coals.
In this study, experiments on both outburst-prone and outburst-resistant/neutral coals are conducted under quasi-static and impact loading using the MTS815 and SHPB systems, respectively. The mechanical properties of both outburst-prone and outburst-resistant coals, including uniaxial compressive strength, indirect tensile strength, elastic modulus, peak strain, and dynamic strength increase factor are compared with strain rate and between coal types. The energy dissipation ratio is proposed to quantitatively describe the energy dissipation of coal responding to the incident energy. These provide the scientific underpinnings for causal mechanisms for the triggering and progress of outbursts.

OBSERVATIONS
In this section, we investigate the principles and processes of experimental observations on both outburst-prone and outburst-resistant/neutral coals in detail to identify their different quasi-static and dynamic failure characteristics. These include aspects of specimen preparation, composition of the SHPB system, theory and basic principles of SHPB tests, and experimental procedures.

| Specimen procurement and preparation
We collected large fresh coal block samples from the 4th coal seam of the Xintian coal mine in Guizhou Province, China. Due to the large-scale coal and gas outburst accident occurred in 4th coal seam of the Xintian coal mine on 5 October 2014, the coal is generally considered as outburstprone coal. The coal is characterized by high in situ gas pressure (p g = 2.5 MPa), small hardiness coefficient (f = 0.5), and low permeability (k = 0.0218 mD). The gas adsorption constant is 26.1 m 3 /t, and the initial velocity of gas emission is 11.2 mmHg. As measured by proximate analysis, the Xintian coal is anthracite with a vitrinite reflectance of 2.13%, moisture content of 1.91%, ash content of 23.28%, volatile matter of 8.81%, and fixed carbon of 66%.
As a comparison, we also collected coal block samples from the 11th coal seam of the Xinzhouyao coal mine in Shanxi Province, China. There is no coal and gas outburst accident happened in Xinzhouyao coal mine during the mining history. This coal is generally an outburst-resistant coal and is characterized by low gas pressure (p g = 0.23 MPa), large hardiness coefficient (f = 4.4), and high permeability (k = 0.756 mD). The gas adsorption constant and the initial velocity of Xinzhouyao coal were not measured. Meanwhile, the Xinzhouyao coal is bituminous coals with a vitrinite reflectance of 0.85%, moisture content of 5.83%, ash content of 14.35%, volatile matter of 2.32%, and fixed carbon of 57.5%. Compared with the Xinzhouyao coal, the Xintian coal is featured by greater vitrinite component and higher metamorphic degree.
The coal blocks were wrapped immediately after cutting and recovery to prevent oxidation and drying during transportation. According to the ISRM suggested methods, specimens with a slenderness ratio of L s /D s = 1:1 (50 mm × 50 mm) and a slenderness ratio of L s /D s = 0.5:1 (25 mm × 50 mm) were drilled from the block samples for dynamic uniaxial compressive strength tests and indirect tensile strength tests (Brazilian split), respectively. 60,61 The two ends of the cylindrical specimens were cut and ground to ensure flatness ±0.05 mm and parallelism ±0.02 mm, and for use in the SHPB tests. A subset of the prepared coal specimens are shown in Figure 2.
The quasi-static mechanical properties of the coal specimens were measured by an MTS815 hydraulic servo-controlled testing system with all measurements made at room temperature. The capacity of the load frame is 100 kN, and displacement control of 1.0 mm/min was used for the uniaxial compressive tests and 0.2 mm/min for the Brazilian split tensile tests. The results of these tests are shown in Figure  3. It can be calculated that the quasi-static compressive strength of Xintian (prone) and Xinzhouyao (resistant) coals are 4.76 and 11.45 MPa, respectively. The quasi-static tensile strength of Xintian and Xinzhouyao coals are 0.68 and 1.45 MPa, respectively. Above information illustrates that the quasi-static strength of Xintian coal is much weaker than that of Xinzhouyao coal.

| Experimental apparatus of SHPB and its basic principles
Dynamic compressive and tensile strength tests on the outburst-prone and outburst-resistant coals were conducted using the SHPB system at the State Key Laboratory for GeoMechanics and Deep Underground Engineering. A schematic diagram and photograph of the SHPB system are shown in Figure 4. The SHPB system consists of a launching device, pressure bars, an energy absorption device, and a signal acquisition and processing system. The launching device comprises a high pressure gas cylinder and gas gun. The pressure bar includes a striker bar, an incident bar, and a transmitting bar. The energy absorption device includes an absorbing bar and a deceleration unit. Finally, the signal acquisition and processing system includes strain gauges and a data acquisition device, and a data processing device to filter noise and analyze the results.
To more accurately obtain the dynamic properties of the coal specimens, a square rubber sheet is used as a pulse shaper. This transforms the incident stress wave from a rectangular wave to an approximately sinusoidal form. Vaseline is applied to both ends of the coal specimens to minimize transverse strain caused by the stress wave. As shown in Figure 5, significant lateral vibrations do not occur during the stress wave propagation. This implies that the SHPB system propagates an essentially one-dimensional wave. Additionally, the strain of the transmitted wave is basically equal to that of the sum of the incident and reflected waves. Thus, the SHPB apparatus also satisfies conditions of stress homogeneity. The rising edge of the incident wave reaches 90 s, which provides sufficient time for the coal specimen to achieve a homogeneous impact stress. Force equilibrium is achieved by preparing specimens with a small slenderness ratio and in maintaining good contact between pressure bars and coal specimens. Slenderness ratios of L s /D s = 1:1 and L s /D s = 0.5:1 were selected for compressive and tensile SHPB tests, respectively. In the SHPB tests, coal specimens are placed between the incident bar and the transmitting bar. The striker bar is launched by a compressed gas gun and collides with the incident bar. The collision induces a longitudinal elastic compressive stress wave (incident compressive pulse, ε i ) in the incident bar. The stress wave then propagates through the incident bar and impacts the coal specimen causing a high rate of deformation. 51 The strain rate delivered to the coal specimen is controlled by varying the striking velocity. When the stress wave propagates to the interface between the coal specimen and the incident bar, it is reflected by and transmitted to the incident and transmitting bars. Parts of the incident pulse are reflected back into the incident bar as a reflected tensile pulse, ε r , while others travel through the specimen into the transmission bar as a transmitted compressive pulse, ε t . Strain gauges on the incident bar and transmitted bar measure the pulse signals and record them to the data acquisition device.
Wave theory is adopted to express the stress-strain curve of the coal material 37 : where γ is strain rate; E b is elastic modulus of the pressure bar, GPa; C b is the velocity of the stress wave in the pressure bar, m/s; A b is the cross-sectional area of the pressure bar, m 2 ; A s is the original cross-sectional area of the coal specimen, m 2 ; and L s is the length of coal specimen, m.
The impact stresses on both ends of the coal specimen are calculated as 54 : Based on the stress homogeneity hypothesis, the impact stress balance on both ends of the coal specimen is achieved by satisfying P 1 (t) = P 2 (t). The above equation can be rewritten as 62 : By submitting Equation (3) into Equation (1), the strain rate, dynamic strain, and dynamic stress are recovered as: Finally, the stress-strain curves of coal specimens under impact loads can be obtained by calculating the measured signals of the reflected and transmitted pulses using Equation (4).

| Experimental procedure of SHPB test
To identify the contrasting dynamic failure characteristics separating outburst-prone from outburst-resistant coals, and to better understand the dynamic failure process, a number of uniaxial compressive SHPB tests and indirect tensile SHPB tests were completed.
For the compressive SHPB tests, a group of 14 coal specimens (Y1-Y14) from the Xintian (outburst-prone) coal mine were loaded to failure, with measured average strain rates varying from 17.18/s to 110.73/s; as a comparison, another group of 7 coal specimens (X1-X7) from the Xinzhouyao (outburst-resistant) coal mine were also conducted, with the measured average strain rates varying from 22.76/s to 105.54/s. A further group of 12 coal specimens (Y15-Y26) from the Xintian coal mine and 6 coal specimens (X8-X13) from the Xinzhouyao coal mine were selected to conduct tensile SHPB tests, with the measured average strain rates varying from 17.25/s to 89.89/s and from 15.62/s to 96.27/s, respectively.
The strain rate in the coal specimen is largely dependent on the striker bar velocity. The relationship between the average strain rate and striker bar velocity for the compressive and tensile SHPB tests is shown in Figure 6. At high speed, the incident bar delivers large kinetic energy. When the incident bar impacts the coal specimen, stress wave propagates and forces both rapid deformation and a high strain rate. The average strain rate of the coal specimen increases with an increase in the striker bar velocity. A least squares fit of the data yields the exponential relation for the SHPB tests.

| RESULTS AND DISCUSSION
We explore the impact of strain rate on the mechanical characteristics of coals-to define key features that control outburst potential, including dynamic deformation moduli, peaks strengths, and the ability to dissipate energy.

| Effect of strain rate on mechanical properties
The dynamic compressive stress-strain curves of the Xinzhouyao (resistant) and Xintian (prone) coal specimens are shown in Figure 7A and 7B, respectively. The dynamic compressive stress paths of both outburst-prone and outburstresistant coal specimens are significantly strain-rate-dependent. For instance, the peak compressive stress (uniaxial compressive strength) of the Xinzhouyao specimen increases from 12. 59 Figure 8A and 8, respectively. Affected by the adhesion of the strain gauge on the incident bar, the incident and reflected pulse signals of coal specimen Y23 were not detected and recordedthere is no dynamic tensile stress-strain curve for this coal specimen in Figure 8B. The strain-rate-dependent response is apparent from the dynamic tensile stress-strain curves. The tensile strength increases with an increase in strain rate. For instance, the peak tensile stress (indirect tensile strength) of the Xinzhouyao specimen increases from 1.67 to 6.91 MPa when the strain rate varies from 15.62/s to 96.27/s, while this value of the Xintian specimen increases from 0.95 to 3.62 MPa when the strain rate varies from 18.04/s to 89.89/s. Compared with Xinzhouyao coal, the Xintian coal has a smaller uniaxial compressive strength and indirect tensile strength, which implies that the outburst-prone coal is usually weaker in both compression and tension than the outburst-resistant coal.
Based on the dynamic stress-strain curves in Figures 7 and 8, we propose an ensemble characteristic response curve to characterize the key features of the stress-strain behavior. The typical dynamic stress-strain curve of coal comprises five stages: compression, linear elastic deformation, microcrack initiation, unstable crack growth, and unloading, 35,39 as shown in Figure 9.
The first stage on stress-strain curve is compression (stage 1). In this stage, the incident and transmitting bars are not completely contacted with the coal sample. The residual space and microfractures within coal sample will be compacted gradually. Accompanying by the closure of microfractures, the deformation resistance of the coal sample increases macroscopically. The second stage is the linear elastic deformation (stage 2), showing a linear increase trend. External loads are not sufficient to accelerate fracture growth or to generate new fractures within the coal sample. This can only result in stable reversible deformations. In this stage, the elastic state of the coal sample remains unchanged, and the elastic energy is accumulating. The slope of the stress-strain curve keeps constant, which represents the Young's modulus. The stage followed by linear elastic deformation is the microfracture initiation (stage 3). The increasing trend of stress slows down with the coal strain, showing an upward convex. In this stage, the microfractures within the coal sample expand and produce new fractures. Plastic deformation gradually dominates as a deformation mode. After that, the coal sample enters the unstable fracture growth stage (stage 4). The fractures expand quickly, and many newly generated fractures connect with the primary fractures caused by the accumulated energy release. On point D, the peak stress achieves and the slope of the curve approaches zero. This point corresponds to the peak strain. The final stage is characterized by the rapid unloading (stage 5). The stress-strain curve decreases rapidly, indicating the completely rapture of coal sample. The bearing capacity of the sample decreases at full speed. The contact surface between the pressure bar and the coal sample becomes uneven, resulting in a variety of curve shapes during unloading, as manifested by various curves in Figures 7 and 8.
In Figure 7, the peak compressive stress increases with strain rate in both cases. Some characteristic parameters, such as the stress-strain gradient in the linear elastic stage and the peak strain in the unstable crack propagation stage, are diverse for each curve. As strain rate increases, the microfracture evolution and unstable fracture propagation stage (stages 3 and 4) are prolonged in both temporal duration and strain ranges, leading to these two stages becoming more significant at higher strain rates. In other words, with the advantage of spatial extension and temporal duration, the cracks within specimens initiate and propagate more rapidly, inducing more dissipation of accumulated energy during these two stages. Then, the stress-strain curves corresponding to different strain rates enter the rapid unloading stage. When comparing the response of outburst-prone and outburst-resistant coals, the outburst-prone coal is more ductile with wide postpeak region and a more significant phase of rapid unloading. The higher strain rate of the coal specimen will lead to a wider range of plastic strain, as well as greater energy dissipation. The stress-strain curves with high strain rates usually decline slightly after reaching the peak value, then enter a plastic plateau, before finally dropping at the conclusion of the test.
In Figure 8, similar features are observed-the peak tensile stress and peak strain increase with the strain rate. Under the same strain rate, the tensile strength of outburst-prone coal is smaller than that of outburst-resistant coal. Normally, the curve rapidly unloads after reaching a peak tensile stress-indicating that the outburst-resistant coal is more brittle.
We extract key features of the dynamic stress-strain curves (Figures 7 and 8), including peak compressive stress, peak tensile stress, strain at peak stress, elastic modulus, and dissipated energy during the entire process of dynamic loading, and compile and present them in Tables 1 and 2.
The key mechanical properties including peak compressive stress, peak tensile stress, peak strain, and elastic modulus are plotted in Figure 10. The peak compressive stress and the peak tensile stress increase linearly with strain rate; the strain at peak compressive stress increases nonlinearly, while the scattered points representing the evolution of elastic modulus are disordered, indicating that the elastic modulus has no obvious relationship with strain rate.
The dynamic increase factor (DIF) is defined as the ratio of dynamic strength to static strength, which reflects the increase in mechanical strength of coal under impact (strain rates). This parameter includes two indexes, the dynamic increase factor for the compressive strength (DIF c ) and the dynamic increase factor for the tensile strength (DIF t ). These are expressed as: where σ c,d and σ c,s are the dynamic and static uniaxial compressive strengths, respectively; σ t,d and σ t,s are the dynamic and static indirect tensile strengths, respectively.  The dynamic increase factors (DIF c and DIF t ) of outburst-prone coal are greater than those of outburst-resistant coal on the whole, indicating that the strain rate hardening effect of outburst-prone coal is more apparent than outburst-resistant coal. With the increase in impact loading, both dynamic compressive strength and tensile strength of outburst-prone coal rise faster than those of outburst-resistant coal. However, the dynamic strength of outburst-prone coal is still smaller than that of the outburst-resistant coal.

| Effect of strain rate on energy dissipation characteristics
The dynamic effect on energy dissipation is fundamentally reflected by the combined action of stress and strain of the coal specimen. During the entire process of dynamic loading and unloading, the energy carried by incident, reflected, and transmitted waves is expressed as 57,63 : where E i , E r , and E t are the incident energy, reflected energy, and transmitted energy, respectively, kJ; σ i , σ r , and σ t are the stresses corresponding to the incident, reflected, and transmitted waves in the pressure bar, respectively, MPa.
The dissipated energy of the coal specimen is given by 63 : The energy dissipation coefficient (C d ), defined as the ratio of dissipated energy to incident energy, is given by:  Figure 12 shows the energy dissipation characteristics of both outburst-prone and outburst-resistant coals in the uniaxial compressive SHPB tests. The dissipated energy increases exponentially with the impact time. Initially, the coal specimen is elastic, and the amount of dissipated energy is small. As impact time increases, the specimen gradually becomes plastic (stages 3, 4, and 5) with the continuous generation of cracks, and their accumulation and expansion-the dissipated energy sharply increases. The faster the impact speed, the larger the strain rate, and the shorter the time taken for the destruction of the coal. Comparing Figure 12A, the energy dissipation of the Xintian (prone) and Xinzhouyao (resistant) specimens appears similar-the dissipated energy increases with an increase in the strain rate. Nevertheless, the dissipated energy of Xinzhouyao specimen (resistant) is larger than that of the Xintian specimen (prone). As shown in Figure 12B, the dissipated energy of the Xintian specimen ranges from 1.106 to 12.279 kJ when the strain rate varies from 17.18/s to 110.73/s, while the dissipated energy of the Xinzhouyao specimen ranges from 3.323 to 23.076 kJ when the strain rate varies from 22.67/s to 105.54/s. The dissipated energy of both Xinzhouyao and Xintian specimens exponentially increases with the strain rate, which are respectively expressed as: Figure 12C shows the variation of incident energy, reflected energy, transmitted energy, and dissipated energy during the entire dynamic failure process of the coal specimens. For the same coal specimen at a specific strain rate, the incident energy > reflected energy > dissipated energy > transmitted energy. Meanwhile, these values of the Xinzhouyao specimen (resistant) are greater than those of Xintian specimen (prone). As shown in Figure 12D, the energy dissipation coefficient of the Xintian specimen (prone) is much smaller than that of Xinzhouyao specimen (resistant) under compressive impact loads. The energy dissipation coefficient increases linearly with strain rate, and their relationships are expressed as: Figure 13 shows the energy dissipation characteristics of the indirect tensile SHPB tests. Also, the dissipation energy increases exponentially with the impact time. The larger strain rate will lead to a shorter duration of coal dynamic failure. The dissipated energy of the Xinzhouyao specimen (resistant) is larger than that of the Xintian specimen (prone). The dissipated energy of the Xintian specimen ranges from 0.061 to 0.373 kJ when the strain rate varies from 17.25/s to 89.89/s, and the dissipated energy of the Xinzhouyao specimen ranges from 0.128 to 1.299 kJ when the strain rate varies from 15.62/s to 96.27/s. The dissipated energy of both Xinzhouyao and Xintian specimens increases exponentially with strain rate, which can be respectively expressed as: For a specific strain rate, the incident energy, reflected energy, dissipated energy, and transmitted energy of the Xinzhouyao specimen (resistant) over the entire dynamic failure process are greater than those of the Xintian specimen (prone). The energy dissipation coefficient of the Xintian specimen is also smaller than that of the Xinzhouyao specimen under tensile impact loads. The energy dissipation coefficient of both Xintian and Xinzhouyao specimens increases linearly with strain rate: It is apparent from the above that the dissipated energy of outburst-prone coal during the dynamic failure process is less than that of outburst-resistant coal under the same degree of impact loading. This may be the reason why the compressive strength and tensile strengths of outburst-prone coal are smaller, and why outburst-prone coal is easier to rupture.

| Discussion on the dynamic failure characteristics of gas outburst-prone coal
Coal and gas outburst is a kind of dynamic disaster occurred during underground mining. Large masses of coals or volumes of gases are suddenly ejected or expelled into working space in few seconds. It takes only a certain milliseconds or microseconds for the outburst coal to be damaged from initial compression. The dynamic characteristics of energy dissipation and mechanical properties of coal may provide the scientific underpinnings for causal mechanisms of outbursts. Many scholars have investigated the dynamic properties by carrying extensive suites of dynamic SHPB tests on rock and common coal materials. However, fewer tests have been completed to identify the dynamic properties on both outburst-prone and outburst-resistant coals. 59 The distinguishment of dynamic characteristics between prone and resistant coals has not been reported. In this study, we conducted experiments on outburst-prone coals against a control sample of outburst-resistant coals under impact loading. The dynamic properties of dynamically failed coal, including uniaxial compressive strength, indirect tensile strength, elastic modulus, peak strain, and dynamic increase factor, together with energy dissipation, were recovered and compared between coal types. Both similar and distinct features of outburstprone and outburst-resistant coals were obtained.
The mechanical properties exhibited similar change trends: The compressive and tensile strengths increased linearly with the applied strain rate, and the peak strain nonlinearly increased with the strain rate. Meanwhile, most of the dynamic stress-strain curves comprised stages of compression, linear elastic deformation, microfracture evolution, then followed by unstable fracture propagation culminating in rapid unloading. The shape of these curves was commonly occurred in many rocks. For instance, Li et al 51  Under the similar impact loading, the tensile or compressive strength of outburst-prone coals is lower than that of outburst-resistant coals. The dynamic increase factor of outburst-prone coals is greater than that of outburst-resistant coals, indicating that the hardening effect of strain rate on outburst-prone coal is more apparent than on outburst-resistant coal. This phenomenon usually occurs in the impact tests on both coals and rocks. Chakraborty 37 investigated three Himalayan rocks (dolomite, quartzite, and limestone) under the same impact loading and found the dynamic increase factor of soft rocks was greater than that of hard rocks. Zhang 59 obtained a similar predicate after series of SHPB tests on coals. Although outburst-prone coal had greater dynamic increase factor than resistant coal, its dynamic strength is still lower than that of outburst-resistant coals due to its much lower quasi-static strength. Meanwhile, the dissipated energy of outburst-prone coal is smaller than that of outburst-resistant coal.
Overall, the outburst-prone coals are commonly characterized by low mechanical strength and high deformability, and small energy dissipation when loaded dynamically to failure. The dynamic failure of outburst-prone coal likely promotes rapid gas desorption into fractures from the coal matrix, and may accelerate the free gas carrying potential for the pulverized coal to enter the mining space and to trigger and participate in outbursts. The fragment distribution and microstructure evolution of dynamically crushed coal were not studied in this paper. These works will be done in the future researches.

| CONCLUSIONS
In this paper, both dynamic uniaxial compressive and indirect tensile tests are conducted on outburst-prone and outburstresistant coals using a split Hopkinson pressure bar (SHPB) system. A full suite of dynamic failure characteristics are comparatively analyzed to contrast the response of outburstprone coals against a control sample of outburst-resistant coals. These characterizations include the dynamic strength, failure processes, and energy dissipation. The following conclusions can be drawn: 1. The dynamic stress-strain response of both outburst-prone and outburst-resistant coal specimens primarily comprises five stages. In the first compression stage, the residual space and microfractures are compacted gradually. In the linear elastic deformation stage, the external loads are not sufficient to accelerate fracture growth, resulting in stable reversible deformations. Followed by the microfracture initiation stage, the microfractures expand to produce new fractures, and the plastic deformation dominates. In the final stage, the fractures expand quickly and connected with each other. When the peak stress achieves, the coal becomes completely raptured. 2. The compressive strength and tensile strength increase linearly with the applied strain rate, while the elastic modulus shows no clear trend with strain rate. Due to a greater dynamic increase factor (DIF), the strain rate hardening effect of outburst-prone coal is more obvious than that of outburst-resistant coal. However, the dynamic strength of outburst-prone coal is still smaller because of its lower quasi-static strength. 3. The dynamic energy dissipation of both outburst-prone and outburst-resistant coals exhibits similar trends. The higher the applied strain rate, the shorter the time required for full destruction of the coal. The dissipated energy increases exponentially with strain rate, while the energy dissipation coefficient increases only linearly. Under identical impact loading, the dissipated energy of outburst-resistant coal is larger than that of outburstprone coal.