*2.1. Sample Specification and AE Sensor*

Samples were taken from working face 8105 of the Tashan Mine, operated by Datong Coal Industry Group Co., Ltd. The coal seam has an average thickness of 18.17 m. The coal was processed into cylindrical samples with a height of 100 mm and a diameter of 50 mm. Coal samples are relatively difficult to process, so there is always a certain amount of error between the actual and ideal samples. The statistical analysis of the 9 coal samples used in the test showed diameters in the range of 49.90 to 50.42 mm, and heights in a range of 98.53 to 100.46 mm. Some of the processed samples are shown in Figure 1.

**Figure 1.** Coal samples.

Mechanical loading was performed using an MTS815 Flex test GT concrete and rock mechanics test system. An AE system (PCI-2, Physical Acoustics Co./PAC) was used for data acquisition. The test setup was exhibited in Figure 2. As shown in Figure 3, eight AE sensors were evenly placed on the outer wall surface of the triaxial pressure chamber (an upper and lower circle of four sensors each). The sensors were coupled with the wall of the pressure chamber by coating their contact area with Vaseline and fixing them using a rubber ring, and were used for real-time monitoring of the time parameters and spatial location of the AE events. The AE sensor model was a Nano30 resonator sensor, with a peak frequency of 300 kHz and a working frequency band of 100 to 500 kHz. Because the sensor was small (see Figure 3) and the chip was at the center of the probe, Vaseline was used as an adhesive to ensure that the chip of the sensor probe fully contacted the measured point, thus ensuring that AE signals could be received. In the test, the wave velocity was 1500 m/s and the threshold was 40 dB.

(a) The mechanical loading system (b) AE system

**Figure 3.** Distributions and dimensions of an AE sensor.

#### *2.2. Test Process*

According to various data collected from field tests, Xie et al. [12,35] summarized the variation characteristics of the abutment pressure of the coal mass in front of the working face induced by actual mining activities. The stress concentration coefficient α under different mining layouts differed under the condition of similar roof property and mining depth, which were 2.5–3.0, 2.0–2.5, and 1.5–2.0 respectively, in the order of nonpillar mining, top-coal caving and protective seam mining. Nonpillar mining is a longwall mining process in which there is no mining roadway and no reserves of coal pillars, while protective seam mining is characterized by mining the protective layer in advance, thus disturbing the adjacent coal seam to weaken its risk of coal and gas outburst. In the simulation experiment, it was assumed that α was 3.0, 2.5 and 2.0 for nonpillar mining, top-caving mining and protected seam mining, respectively. Therefore, based on the theory proposed by Xie et al. [12,36], the entire mining-induced mechanical loading (unloading) process can be divided into three stages: the hydrostatic pressure stage (OA), the first unloading stage (AB) and the second unloading stage (BC, BD, or BE), as shown in Figure 4. In Figure 4, γ refers to the average specific gravity of rocks on top of the coal seam and fc' is the triaxial compressive strength of coal. The simulated mining depth was approximately 1000 m and the hydrostatic confining pressure in OA was defined as 25 MPa; from this, the horizontal and vertical stress in each stage can be calculated.

Figure 5 exhibits the loading (unloading) scheme corresponding to different mining layouts at 1000 m depth. In the first stage, to simulate the influence of initial hydrostatic pressure on coal rock before mining, the hydrostatic confining pressure was applied to an in situ stress state at point A (σ<sup>1</sup> = σ<sup>3</sup> = γ*H* = 25 MPa) at the same confining pressure loading rate of 3 MPa/min under each of the three mining layouts. The second stage was the first unloading stage. With the occurrence of excavation disturbance, the axial stress of coal rock increased and the horizontal stress decreased. All three groups of coal samples were unloaded, starting from point A, at a confining pressure unloading rate of 1 MPa/min; at the same time, all were loaded to point B (σ<sup>1</sup> = 1.5 γ*H* = 37.5 MPa, σ*<sup>3</sup>* = 0.6 γ*H* = 15 MPa) at an axial pressure loading rate of 2.25 MPa/min. In other words, the axial stress was increased at a rate of 2.25:1 and the lateral stress was reduced. After point B, the specimen was divided into three loading paths to reflect the stress effect produced by different mining methods. The three mining layouts were associated with the same confining pressure unloading rate of 1 MPa/min and axial loading rates of 2.25 MPa/min (protective seam), 3.5 MPa/min (top-coal caving) and 4.75 MPa min (nonpillar) until sample failure.

**Figure 4.** The stress state of coal mass in front of the working faces under three different mining layouts.

**Figure 5.** The loading (unloading) scheme under different mining layouts (γ*H* = 25 MPa).

#### **3. Results**
