Kinetic characteristics of coal gas desorption based on the pulsating injection

International Journal of Mining Science and Technology 24(2014)631-636Contents lists available at Science DirectInternational Journal of Mining Science and TechnologyELSEVIERurnalhomepagewww.elsevier.com/locate/ijmstKinetic characteristics of coal gas desorption based on the pulsatingInjectionNi Guanhua, D, * Lin Baiquan Zhai Cheng, Li Quangui a, D, Peng Shen. Li Xianzhong CSchool of Safety Engineering, China University of Mining 8 Technology, XuzhouState Key laboratory of Coal Resources and Mine Safety, Xuzhou 221116, ChinaSchool of Energy Science and Engineering, Henan Polytechnic University, Jiaozuo 454000, ChinaARTICLE INFOABSTRACTArticle histoorder to understand the kinetic characteristics of coal gas desorption based on the pulsating injectionReceived 13 October 2013(PI), the research experimentally studied the kinetic process of methane desorption in terms of the pl andReceived in revised form 9 January 2014Accepted 18 March 2014hydrostatic injection(HI). The results show that the kinetic curves of methane desorption based on Pl andAvailable online 29 August 2014HI are consistent with each other, and the diffusion model can best describe the characteristics of methane desorption. Initial velocity, diffusion capacity and ultimate desorption amount of methane desorptionafter Pl are greater than those after HI, and the ultimate desorption amount increases by 16.7-39.7%Methane decay rate over the time is less than that of the Hl. The Pl influences the diffusion model paramHydrostatic injectioneters, and it makes the mass transfer Biot number b, decrease and the mass transfer Fourier series FoMethane desorptionincrease. As a result pi makes the methane diffusion resistance in the coal smaller methane diffusion rategreater, mass transfer velocity faster and the disturbance range of methane concentration wider than HITherefore, the effect of methane desorption based on Pl is better than that of HIo 2014 Published by Elsevier B V on behalf of China University of Mining Technology1 Introductionshow that a large amount of gas was expelled during the closureof fracture due to the confining pressureThe methane is not only a kind of clean energy and valuableWith the increase of coal mining depth and exploitation intenchemical material, but also a substantial root causing coal and sity, coal outburst and gas explosion seriously threaten minersgas outburst as well as explosion. In the natural coal, the fracture personal safety, and restrict the development of coal mines. Effinetwork is arranged by a large number of pores and cracks, from ciently and safely relieving pressure of coal seam is an importantthe defect in a molecular or atomic scale to visible cracks, with guarantee to the safe production of coal mines. In recent years,considerable specific surface area and strong methane adsorption all kinds of hydraulic measures have been taken in managing minecapacity. The methane is adsorbed on the inner surface of the coal gas and prevention of coal and gas outburst by coal mining entermatrix physically. Methane adsorption and desorption characteris- prises and scholars, including hydraulic fracturing, coal seam watertics are related to not only the pore structure of coal, metamor7. Conventional hydro-hism, ash, moisture and other intrinsic factors, but also gas static injection(Hi) can play a role in wetting the coal. Howeverpressure, temperature, stress, electric field, magnetic field and due to blocking up the channel of the gas desorption and migrationacoustic field [1, 2]. Chen et al. determined the methane adsorption by the water, there is a blocking effect on the gas desorption, whichcharacteristics of the anthracite at different temperatures by the has been confirmed by many scholars [8, 9 Pulsating injection(pl)olumetric method [3]. Their results show that the equilibrium is a new technology based on the original coal seam water injectime of methane adsorption increased from 5 h at 50C to 25 h tion and aimed at improving pressure relief of the coal seamdied by He et al, using an experimental seam permeability (10 Zhai et al. took hydraulic fracturing expesystem with temperature-pressure coupling [4 Their results iments and mechanical analysis under the different pressures andfrequencies, and analyzed the characteristics of coal fatigue damage [11]. Their results showed that alternating stress was formedhor at: School of Safety Engineering. China University ofXuzhou221116, China.Tel:+8615950668058.at the end of coal中国煤化工 he pulsating water,gh0101@163.com(g.Ni).and fatigue damag-le to the repeatedCNMHGhttp://dx.doiorg/10.1016/j.ijmst.2014.03.0232095-2686 2014 Published by Elsevier B V on behalf of China University of Mining TechnologyG Ni et al. / International Journal of Mining Science and Technology 24(2014)631-636Decompression valveManometer地中十Shutoff valve 3High-pressure gasPulse injectiShutoff valve 5 Sydrostatic injectiorpumpFig 1. Methane desorption system based on the Pl and HItime after PI and Hl, and comparatively analyzes the kinetic char4-4acteristics of coal gas desorption based on Pl2. Kinetic characteristics of methane desorption-e Pl of 3 MPa -o-HI of 3 MPaIn order to explore the kinetic characteristics of methane4 PI of I mpa亠 HI of I MPadesorption, this research has made unremitting efforts to establish20040060080010001200a variety of dynamic models, and put forward a number of equations to calculate the changes of methane desorption amount overthe time in coalFig. 2. Curves of gas desorption amount with time based on the Pl and HI2.1. Empirical equationsaction of compression and expansion. Li et al. investigated the gen-Wang and Yang determined the velocity of the coal gas desorperation and propagation of pulsating stress wave and the mecha- tion using the gravimetric method, and put forward Eq. (1)to indilaboratory experiment and field test [12 ] Their study shows that cate the amount of desorption over the time [15]reflection and superposition of pulsating stress wave could makestress increase in some areas, and make the stress decrease in other )1+Btareas. Li et al. took the experiments in three aspects including dif-ferent water injection mode and pulsating frequency and parame- where Q(t) is the desorption amount a the saturated desorptionter combination by pulsating hydraulic fracturing system [13]amount;B the desorption constant; and t the time.They concluded a result that the pulsating pressure should be controlled properly so that it could provide sufficient time for the pul- 2.2. Diffusion modelsating water in the coal when the pulsating hydraulic fracturingas implemented. Ni et al. determined and analyzed the sealedOne, study makes the assumptions as follows: Firstly, this studyparameters of pulsating hydraulic fracturing using the diameters makes the assumptions as follows: ( 1)coal particles were spheri-of 95 and 108 mm rough inner wall seamless steel pipes as simu-re holation drills [14]. They found that the reasonable sealed length was and (3)methane flow in coal particles followed the mass conserva13 m when the pulsating hydraulic pressure was 16 MPa. However, tion law and the principle of continuitythe impact of the Pi on the kinetic characteristics of gas desorptionUnder the above assumptions, the diffusion coefficient and theis still not clear, and there is no relevant study to date. Therefore, coordinate were independent. ignoring the influences of concenthis paper experimentally studies the gas desorption changes with tration C and time t on diffusion coefficient and taking polarGas pressure(MPa) Injection modeB(min) B(min Qx(mL/g) b(mrrelation coefficient rEmpirical formula 37.71491.11820.16936.4193.8843(b)Diffusion model3PI0.00550.2868099178.51730.00425.54570.9644TYH中国煤化工CNMHGG. Ni et al. / International Journal of Mining Science and Technology 24 (2014)631-636一 Measured valuePenetration model fittingDiffusion model fittingGas pressure of 3 MPa(b) Gas pressure of I MPaFig 3. Measured values, empirical formula, diffusion model and penetration model fitting of gas desorption based on the Pl.ordinate, it can be obtained the Fick,s second law(Eq (2))under the vacuum pump the coal samples were degassed 6 h at roomball coordinate systemtemperature then shutoff valve 1 was opened and shutoff valves2, 3, 4 and 5 were still closed The high-pressure methane wasac/a2c 2 aar(2) injected into the desorption cylinder. After 24 h, the methaneadsorption of coal samples reached equilibrium, then the methanedesorption experiments could be carried out. Drainage methodwhere C is the methane concentration; D the diffusion coefficient, r was adopted to determine the desorption amount, and desorptionthe radius of coal particle; and t the timeYang and Wang has gained the approximate solution of meth- experiment. If using the PI, shutoff valves 1, 2, 3 and 5 should beane diffusion equation(Eq(3)by solving Ficks second law(Eq. closed, and shutoff valve 4 should be opened. Then the pulse injec(2))and computer numerical fitting under the spherical coordiion pump was opened and the pulsating water was injected intonates, using the heat conduction equation and mathematical metthe coal samples for 10 min. If using the Hl, shutoff valves 1, 2, 3ods [16and 4 were closed, and shutoff valve 5 was openedQ(t)=Q、√1(3) hydrostatic injection pump was opened and the hydrostatic waterwas injected into the coal samples for 10 minwhere b is the parameter, and b=4T'Dld2 d the diffusion coefficient; d the radius of coal particle; Qo the ultimate desorption 3.2. Experimental resultsmount; and K the correction factor. when B-value ranges between65797×10-3-6.5797×106,theK- value is taken as0.96the experiments, we tested the methane desorption curvesbased on the pI and HI under the gas pressure of 1 and 3 MPa23. Permeation modelrespectively at room temperature. The pulsating frequency was25 HZ, and both pressures of pulsating water and hydrostatic waterIn Eq (4), the exponential form was widely used to express the were 3 MPa. Fig. 2 shows the experimental curveschange of the coal seam pre-pumping methane over the time [171The methane desorption kinetic curves are consistent betweenQ(t)=Q、[1-eXp(-bt)PI and Hi(Fig. 2). The initial speed of methane desorption andthe ultimate methane desorption amount increase with gas pres-where b is the attenuation coefficient of the diffusing speed over the sure and methane desorption speed based on the PI in coal is sigtimenificantly greater than that of Hl. In the initial stage the methane is3. Experimentalane desorption amount tends to balance after 200 min. Moreoverthe ultimate desorption amount after the pl is greater than that3. 1. Experimental apparatus and methodof HI ( Fig. 2)The self-designed methane desorption system based on the pl 3.3. Analysis of experimental resultsand HI was used as the experimental equipment, including waternjection system, vacuum pumping system, gas supply systemAccording to data in Fig. 2, the kinetic parameters of methaneand test system. Water injection system includes the pulse injec desorption of empirical formula, diffusion model and penetrationtion pump and hydrostatic injection pump. Pulsating water was model can be calculated through regression analysis table 1)supplied by pulse injection pump, which pulsating intensity andEmpirical formula is not used to describe the characteristics ofquency were 0-25 MPa and 25 HZ, respectively. Test system methane desorption based on PI and HI in coal, with a correlationmainly consists of desorption cylinder, manometer and graduated coefficient(R)of around 0. 1-0.2 Table 1). However, the diffusionlinder. Coal samples in the desorption cylinder were degassed model and the penetration model can be in a good description forthrough the vacuum pump. The methane adsorption of coal sam- it, with the correlation coefficient(r)close to 1. In Pl, the kineticdangerous zone of Songshu Coal Mine. The coal particles at 40- b That is to sa d b increase with gas pressure. In the same gasdecompression valve.essure, the methane desorption kinetic parameters b and QThe coal samples were collected from the coal and gas outburst after the pl are中国煤化工 on contrary foracity and ultimate60 meshes were filtrated as experimental coal samples, and 120g desorption amountCNMH GI and it is on concoal particles were put in the desorption cylinder. As shown in trary for methaneLuc Llc. IvIUI over. the ultimateFig. 1, shutoff valve 2 is opened, while shutoff valves 1, 3, 4 and desorption amount based on the pi increases by 16.7-39 7% com5 are closed at the beginning of the experiment. By turning on pared with the hi in the same gas pressure (table 1)34G Ni et al /International Journal of Mining Science and Technology 24(2014)631-636Time (min)Time(min)0020030400500600ln1-Q,/Q)=00053+ln06702c三5ln(1-Q/Q.)=0006+m08R2=0.9725(a)PI/gas pressure of 3 MP(b)HI/gas pressure of 3 MPaTime(min)00400600800100012000200400600800100012001400(1-Q/Q)=0.0061+nnl-Q/Q.)=00028+ln-3.0(c)PIgasI MPaof 1 MPaFig 4. Experimental data and the regression curves of gas diffusion parameters based on the Pl and HIbased on pl and HGas pressure(MPa)Injection moded(minx(10-m/sD(1013m2/sCorrelation coefficient R20.00760.81575625100330.97250.676256.28480967956.23880.55708568Fig 3 shows the methane desorption curves of measured value, where c is the concentration of methane d the diffusion coefficient;empirical formula fitting, diffusion model fitting, and penetration r the distance of the center of the coal particle to one point; Co thenodel fitting. The diffusion model fitting effect of kinetic charac- initial equilibrium concentration; ro the initial radius of coal parti-teristics of methane desorption is best in both initial and equilib- cle; a and b Langmuir constants: po the initial equilibrium pressurerium phases: the penetration model fitting effect follows by a the mass transfer coefficient between coal surface methane andDiffusion Model, which has good effect in the initial phase, but less free methane; and cr the free methane concentration in coaleffective fitting in equilibrium phase: the empirical formula is fractureworst, which has worst fitting effect in both initial and equilibriumAccording to Nie, the formula t>0, AfI-o=0 should bephases( Fig 3). Therefore, the diffusion model can best describe the replaced by t>0, ar Ir-0=0, physical meaning is that the concen-characteristics of the methane desorptiontration gradient at the particle center(r=0)is 0 at any moment,and it is the partial derivative of c to the r rather than the t [184. Kinetic characteristics of methane desorption based on the pl Otherwise, the concentration c of particles at the center is a con-stant, which does not meet the actual gas diffusion. Thereforebest describe the characteristics of methane desorption in coal. coordinates is wrong in his paper, and this article s under sphericalthat of the HI, because the Pl affects the kinetic parameters of thehe eq.(5) is the secondary parabolicdiffusion modelsolved by the method of separation variables. The solution isThe kinetic model of diffusion definite solution was given underspherical coordinates in coal by Nie [18Q rD(5+2)1 Ba(B中国煤化工t=0.00.1=0=0where f is the mass0=Dt/r: B1 a solu-Do=a(c-crlrertion of the equation; tan B=B(1-oro/D)=B/(1-B); B the mastransfer Biot number; and B= oro/DG Ni et al. / International Journal of Mining Science and Technology 24(2014)631-63600400600800100012001400(a) Gas pressure of 3 MPiFig. 5. Curves of the mass transfer fourier series Fo with the time.They are the solutions in the form of series, and are a rapidly velocity faster and the disturbance range of methane concentrationregent series when Fo >0. Therefore taking the first term to wider. Therefore the effect of methane desorption after the pl ismeet the engineering precision, then the Eq (6)can be changed as: better than that of HIQt=6(1 cos p1-sin B1)B1(B1-B1 sin B1 COS BI(7) 5. ConclusionsTake logarithm of the Eq (7)on both sides(1) The methane desorption kinetics curves are consistentetween the pi and hl and can be used to describe the dif-A=6sin,),2fusion model and penetration model. The diffusion model(B-B1that describes methane desorption characteristics is theest, and followed by penetration model Empirical formulaBy the knowledge of mass transfer, it is found that the greatedoes not indicate the characteristics of methane desorptionthe mass transfer Fourier series Fo is, the greater the diffusion ratebased on the piof the methane in the object becomes; the faster the mass transfer(2) The initial speed of methane desorption and the ultimatespeed is, the more extensive the concentration disturbance range ismethane desorption amount increase with the increasing[19]. The smaller mass transfer Biot number B, is, the smaller thegas pressure. In the initial stage, the methane is desorbednternal diffusion resistance becomes in objects, the stronger diffusion capacity is, the more consistent the gas diffusion concentraquite fast, and it gradually slows down with time. The methane diffusion capacity and ultimate desorption amount,tion in inside of the objects and the smaller the diffusionwhich increases by 16.7-39.7%, are greater in Pl than thatconcentration difference between the internal and the external isin HL, but they are on contrary for methane diffuse attenuan(1-Q/Q、)=-it+lnAtion speed over time.()The methane desorption capability after the pI is strongerwhere Qr, Qo and t are determined by methane desorption experithan that of Hl, because the pi affects the kinetic parametersment;and 2. and A are obtained by linear regression. Then B1, theof the diffusion model. Compared with the Hl, the Pl makesdiffusion coefficient d and the mass transfer biot number b: canle mass transfer Biot number b decrease a little and thebe calculatedmass transfer Fourier series Fo increase. Then the PI makesAccording to data in Fig. 2, methane diffusion parameters 7. anthe methane diffusion resistance in the coal smaller. methA values based on the Pl and HI under the gas pressure of 1 MPaane diffusion rate greater, mass transfer velocity faster andand 3 MPa can be calculated by regression analysis(Fig. 4). Thenthe disturbance range of methane concentration wider thanB1, the diffusion coefficient D and the mass transfer Biot numberHIB i can be calculated by eqs. (9)and(7), as shown in Table 2Regression curves of methane diffusion parameters can fit theexperimental data based on Pl and HI well, and there are compar- Acknowledgmentsatively accurate methane diffusion parameters of 2 and A valuesFig. 4). The difference of affecting on the mass transfer Biot ntThis work was financially supported by the National Basicber b between the PI and HI is small, and all B, are slightly larger Research Program of China(No 2011CB201205), the National Natthan 56(Table 2). However, the mass transfer Biot number B; after ural Science Foundation of China(No 51274195), the Natural Sci-the Pl is still slightly less than that of the Hl. It illuminates that the ence Foundation of jiangsu Province of China(No. BK2012571), theeffect on internal diffusion resistance in coal from the Pl is small, National Major Scientific Instrument and Equipment Developmentut it is still slightly less than the internal diffusion resistance project of China(No. 2013YQ17046309), and the Education Departbased on the Hl, and diffusion ability is slightly stronger than that ment Science and Technology Key project of Henan Province ofof hiChina( 14B440007), which is greatly acknowledgedThe same gas pressure, gas diffusion coefficient d after the pl isgreater than that of Hl, indicating that the mass transfer Fourierseries Fo increases because of the pulsating effect(Table 2). Fig. 5Referencesshows the curves of the mass transfer Fourier series Fo with the中国煤化工ane and carbon dioxidtime. The pI has a large impact on the mass transfer Fourier seriesGeol2002:51HCNMHGanian coals Int J coalFo(Fig. 5). The mass transfer Fourier series Fo after the PI are allgreater than that of HI at the same time. Compared with the hi,[2] Ni GH, Lin BQ, Zhai C. Influence of water incursion on the pore characteristicsoal from Eastern China at different temperatures. Disaster Adthe Pl makes the methane diffusion rate greater, mass transfe2013:6(11):43-54G Ni et al. / International Journal of Mining Science and Technology 24(2014)631-636[3] Chen CG, Xian XH, Zhang D]. The dependence of temperature to the adsorption [11 Zhai C, Li M, Sun C, Zhang ]G, Yang W, Li QG. Guiding-controlling technology ofcite coal and its char. Coal Conveyor 1995: 18(3)[4 He MC, Wang CG, Li D). Desorption characteristics of adsorbed gas in coal [12 Li XZ, Lin BQ, Zhai C, Ni GH, Li wZ. Relaxation study of cement based groutingial compression. Chin J Rockmagnetic resonance. Int Min Sci TechnolMech Eng2010:29(5):865-72012:22(6):82[5 Lv YC. Application the hydraulic fracturing technology in the high pressure and13 Li QG, Lin BQ, Zhai C, Ni GH, Peng Shen, Sun C, Cheng YY. Variable frequency oflow permeability mine. J Chongqing Univ 2010: 33(7): 104-6I Liu Z, Li ZH, Yang YL Study of outburst prevention technology bySci technol2013:23(6):847-53ive hydraulic drilling downward through coal-seam. J Min Saf [14 Ni GH, Lin BQ, Zhai C, Li xZ, Li QG Borehole sealed parameter measurementnd analysis of pulsating hydraulic fracturing. J China Univ Min Technol[7 Tang JX, Jia JQ, Hu GZ. Design andmentation of high-pressure water jet2013:42(2):177-8215]2007;28(7}:1501-4China Coal Soc 1980: 5(1): 41-53.[17 Yang QL A preliminary study on the change regulation with the time of cinderss roadway caused by coal and gas outburst. Int J Min Sci Te[18 Nie BS, Wang EY, Guo YI Mathematical and physical model of gas diffusion2014;24(1):23hrough coal particles. J Liaoning Tech Univ 1999; 18(6): 582-5[10 Ni GH, Lin BQ, Zhai C Electromagnetic radiation mechanism of pulse hydraulic [19 Gao JR. Momentum, heat, mass transfer principle. Chongqing: Chongqingfracturing and application in mine Safe Coal Min 2011: 42(11): 1-4中国煤化工CNMHG