不同侧压力系数条件下爆破扰动诱发巷道渐进破裂过程.pdf

Trans. Nonferrous Met. Soc. China 302020 2518−2535 Progressive fracture processes around tunnel triggered by blast disturbances under biaxial compression with different lateral pressure coefficients Yi LONG, Jian-po LIU, Gang LEI, Ying-tao SI, Chang-yin ZHANG, Deng-cheng WEI, Hong-xu SHI Key Laboratory of Ministry of Education on Safe Mining of Deep Metal Mines, Northeastern University, Shenyang 110819, China Received 11 January 2020; accepted 22 June 2020 Abstract To investigate the progressive fracture processes around a tunnel triggered by static stress and dynamic disturbance, experiments and numerical simulations were pered. The results show that the spatial distributions of acoustic emission AE events become very different as lateral pressure coefficients change. The combined effect of static stress and dynamic disturbance causes the damage around the tunnel, and initial stress conditions control the damage morphology. The blast disturbance cannot fundamentally change the damaged area but will deepen the extent of damage and accelerate the failure speed. The more significant the difference between the vertical and horizontal stresses is, the higher the impact on the tunnel by the dynamic disturbance is. The AE activity recovers to a relatively stable state within a short time after the blast and cons to power-law characteristics. Key words tunnel damage; blast disturbance; lateral pressure coefficient; acoustic emission; power-law fitting 1 Introduction Under high in-situ stress conditions, in the miningprocess,rockmassesareinevitably subjected to frequent disturbances that can be classified into unloading effects, i.e., the stress change in the surrounding rock mass after the ore is mined, and dynamic disturbances, such as blast. Therefore, the rock mass is subjected to high stress and dynamic disturbance [1]. For high-production mines, especially those that use medium-length hole blast or long-hole blast, one uses more than hundreds of kilograms or even several tons of explosives. Strong blast disturbances can easily lead to spalling and collapse of the surrounding rock of tunnels, even generating dynamic hazards, such as rockbursts[2−5].Statisticaldatashowthat two-thirds of mine rockbursts occur after blasts. For example, blast vibrations damaged more than 20 m of the roadway in the Hongtoushan copper mine, one of the deepest nonferrous metal mines in China, at a mining depth of 1000 m [6]. In the Gujiatai iron mine, a blast disturbance caused the collapse of a large rock mass in a stope, and the production was subsequently suspended for several years, resulting in severe economic losses. During the deep tunnel excavation of the Jinping II hydropower station, a blast disturbance caused a time-delayed rockburst, for which the height, length, and maximum depth of the damage were 4, 30, and 0.9 m, respectively [7]. In deep rock masses, stresses quite possibly concentrate around the tunnel. As a result, the triggering effect of dynamic disturbance for rock Foundation item Project 2017YFC0602904 supported by the National Key Research and Development Program of China; Project 51974059 supported by the National Natural Science Foundation of China; Project N180115010 supported by the Fundamental Research Funds for the Central Universities, China Corresponding author Jian-po LIU; Tel 86-13514265478; E-mail liujianpo DOI 10.1016/S1003-63262065398-5 Yi LONG, et al/Trans. Nonferrous Met. Soc. China 302020 2518−25352519 damage is more prominent under such conditions. At present, the research on the excavation-damage zone EDZ of tunnels is mainly conducted under static loading using rigid or servo-controlled testing machines. However, the damage in rock mass inducedbydynamicdisturbancescannotbe ignored,especiallyunderhighin-situstress conditions [8]. The split Hopkinson pressure bar SHPB is a commonly-used device for studying the characteristics of rock dynamics. With the help of the SHPB, many researchers [9−11] have tested the dynamicstrengthofrocksandsubsequently establishedconstitutiverelationsandproposed many models based on fracture mechanics or damage mechanics. In addition to the SHPB device, several other instruments have been developed to simulate the deation and failure evolution under dynamic disturbance [12−16]. Numerical simulation was also pered by ZHU et al [8,17] and LI et al [18] to uate the dynamic stress concentration around the tunnel, and the areas prone to failure induced by dynamic disturbance were found.However,experimental workson rock damage due to dynamic disturbances are mainly restricted to surface observations, and the evolution characteristics of internal damage still need to be further studied. When a rock is subjected to loading conditions,microcracksgraduallygenerate, propagate and coalesce, accompanied by a release of elastic strain energy, i.e., acoustic emission AE. The AE technique can continuously monitor the temporal-spatial evolution of microcracks, thereby revealing the deation and failure process of rock. A series of achievements have been realized, e.g., the AE location algorithms, the temporal- spatial evolution of microcracks in rock bodies, and the changes in manyAE parameters [19−32]. In this study, cement mortar was used to create specimens with a prefabricated circular hole in the middle. The biaxial stress is applied with different lateral pressure coefficients that are calculated from the stress condition around a tunnel in the Ashele copper mine, and the blast is used to generate dynamic disturbances. In this process, the spatial-temporal evolution of the AE activity was analyzed to reveal the damage characteristics under static stress and blast disturbance conditions. Also, numerical simulations were pered to verify the experimental results, to provide a theoretical basis for support optimization for deep tunnels. 2 Experimental 2.1 Specimen preparation and AE monitoring scheme The specimens used in this work were made of cement mortar of 300 mm 300 mm LW, 250 mm in thickness Fig. 1. The mass ratio of cement to sand and water was 14.10.9. A circular hole with a diameter of 57 mm was drilled in the middle of each specimen. Also, specimens with the same mass ratio were used to measure the uniaxial compressive strength UCS, and the results showed that the UCS of this batch of specimens was approximately 24 MPa. Fig. 1 Specimen structure, AE sensor arrangement, and blast position Seventeen sensors with a response frequency range of 50−400 kHz were arranged on the front and back faces of the specimens, eight on the front face, and nine on the back face. When the sensors were fixed on the specimen faces, rubber bands and vaseline were used for coupling between sensors andspecimen.Asensorhighway-IISH-II, American Physical Acoustics Corp. system was used for the AE monitoring. The system uses an 18-bit A/D switching technology that allows instant time wave recording with a maximum upper limit amplitude of 10 V. The sampling threshold, sampling frequency, and sampling length were set to be 45 dB, 10 MHz and 5120, respectively. 2.2 Blast disturbance scheme The blast position was located in the middle of the specimen, 60 mm away from the bottom face Yi LONG, et al/Trans. Nonferrous Met. Soc. China 302020 2518−25352520 Fig. 1. When the explosive was detonated, the detonation gas and shock waves could quickly induce ruptures around the blast hole that may propagate and extend to the tunnel, thereby causing failure around the tunnel. The explosive was placed in a steel pipe with a bottom closed-end to avoid the blast breaking the specimen. The diameter, thickness, and length of the steel pipe were 8, 0.1, and 125 mm, respectively. The explosives were preparedbymixingcyclonitewithpotassium picrate because the cyclonite has high brisance and low sensitivity, while the potassium picrate has opposite characteristics. The optimized mass ratio of the two materials was 11 as obtained from a preliminary test. Each blast used 1 g of explosive. After the explosives were placed in the steel pipe, fine sand was used to fill the blast hole, and electric blast was adopted. 2.3 Loading path The actual stress condition around the tunnel during the mining process determines the lateral pressure coefficients during biaxial compression. The Ashele copper mine is located in the Xinjiang Uygur autonomous region of China. Currently, the development depth and mining depth are 1200 and 900 m, respectively. The tectonic stress is dominant and increases with the mining depth. According to the stress test results, the horizontal stress is 1.4 times the vertical stress when the mining depth exceeds 800 m from the surface. Under high-stress conditions and blast disturbances, ground hazards, suchasspalling,collapse,androckbursts, frequently occur. Figure 2 shows the vertical stress distribution around the tunnel by the unloading effect of No. 4 stope mining at the 0 m level approximately 900 m in mining depth in the Ashele copper mine. Because the stope is vertical, the unloading effect on the horizontal stress is small. From Fig. 2a, the vertical stress increases from the boundary of the No. 4 stope position P1 to the maximum value position P5 and then gradually returns to the in-situ stress level position P13. Figure 2b shows the change in the lateral pressure coefficient k, i.e., the ratio of the vertical stress to the horizontal stress, within 60 m from the stope boundary. Because the lateral pressure coefficient of the in-situ stress is 1.4, in this study, four lateral pressure coefficients were investigated, i.e., 0.8, 1.0, 1.2, and 1.4. The horizontal stresses σ2 were designed to be 16 MPa, and the vertical stresses σ1 were designed to be 20, 16, 13.3, and 11.4 MPa, as shown in Fig. 3 and Table 1. The experiment is divided into three stages. First, σ3and σ1were simultaneously loaded at 3 MPa/min until the predetermined values were reached. Second, six blast disturbances were carried out, and each blast process was conducted after the AE signals stopped generating. Eventually, the vertical stress σ1or horizontal stress σ2was loaded again until the specimen failed. Fig. 2 Schematic diagram of stress distribution around tunnel induced by unloading effect of deep stope mining in Ashele copper mine a Vertical stress distribution; b Lateral pressure coefficients, k, in surrounding rock mass of tunnel at different distances from stope 3 AE sequence with loading and blast disturbances 3.1 Filtering of blast signals At the time of the blast, the stress wave is transmitted to the specimen through the steel pipe and propagates into the specimen. With the high- energy blast trigger stress waves, the waves may continuetopropagateinthespecimenfor some time, and they may be mixed into the signals Yi LONG, et al/Trans. Nonferrous Met. Soc. China 302020 2518−25352521 Fig. 3 Schematic diagram of loading path and blast scheme a k0.8; b k1.0; c k1.2; d k1.4 Table 1 Loading plan and blast parameters Loading planHorizontal stress, σ2/MPaVertical stress, σ1/MPaBlast numberExplosive mass/gk 11620610.8 21616611.0 31613.3611.2 41611.4611.4 generatedfrommicrocracks.Therefore,these interferencesignalsshouldbefilteredbefore analyzing the AE activity. Automatic s are beneficial to the wave identification [33−35], especiallyforlargeamountsofsignaldata. However, this study pers manual filtering of blast signals because there are only six blasts for each specimen, and the blast time is known. Also, the process of filtering the interference signals duetoblastconsidersthedifferenceofthe parameters between signals induced by blast and microcrack generation and the attenuation law of blast signals. Due to the considerable amount of explosive energy released during detonation, some parameters of the blast signals such as the rising time, hits, energy, amplitude, and duration time, are very high. Especially, the energy and amplitude of the many blast signals that exceed the acquisition range of the AE system reach significant values. For example, the amplitude of the typical blast signal shown in Fig. 4a is so large that a peak clipping phenomenon appears the acquisition range of the Fig. 4 Waves of blast signal a and crack signal b Yi LONG, et al/Trans. Nonferrous Met. Soc. China 302020 2518−25352522 amplitude in this AE system is 10 V. Also, the duration time of this blast signal is very long and does not decay to a stable state within the sampling time. On the contrary, the wave parameters of the signal from the generated microcracks are much lower than those of the blast signal see Fig. 4b. Therefore, the blast signal can be identified and filtered out according to the wave parameters. On the other hand, the blast signal gradually weakens during propagation. When its amplitude is smaller than the sampling threshold, the sensors can not detect it. Many factors affect the signal attenuation, including the excitation conditions, wavefrontdiffusion,absorptionattenuation, interfacial reflection, interfacial transmission, and interfacial morphology. Among them, this study focuses on the absorption attenuation and interfacial reflection. The attenuation ula defined by FUTTERMAN [36] is as follows 1 2 1 ln A xA   1 where α is the attenuation coefficient, ∆x is the propagation distance, A1is the amplitude at the source and A2is the amplitude after the propagation distance, ∆x. For the interfacial reflection factor, the reflection coefficient is related to the wave impedance of the materials on both sides of the interface. The wave impedance at normal incidence can be written as follows 2 21 1 2 21 1 vv R vv      2 where R is the reflection coefficient; ρ1, v1, ρ2and v2 arethedensitiesandthewavepropagation velocities of the specimen and the pressure head, respectively; and ρ1v1and ρ2v2are the wave impedancesofthetwomaterialskg/m2s. According to Eqs. 1 and 2, the time that takes for the blast signal amplitude to decay to the sampling threshold is less than 0.5 ms based on the attenuation coefficient and wave impedances of cement mortar and the wave impedances of steel. In summary, one can identify and remove blast signals by considering their parameter characteristics and attenuation law. 3.2AE sequence characteristics Figure 5 shows the AE sequence during the specimen fracturing with different lateral pressure coefficients. The AE activity closely relates to the stressloadingandblastdisturbance.Forthe specimen with lateral pressure coefficient k0.8 Fig. 5a, AE signals are continuously generated during the first loading phase, indicating that the internal damage of the specimen gradually increases during this phase. Also, the AE activities show a “strong → weak → strong” trend. As some tiny air bubbles were probably mixed into the cement mort