In situ strength of coal bed based on the size effect study on the uniaxial compressive strength

International Journal of Mining Science and Technology 24(2014)74Contents lists available at Science DirectInternational Journal of Mining Science and TechnologyELSEVIERurnalhomepagewww.elsevier.com/locate/ijmstIn situ strength of coal bed based on the size effect study on the uniaxial Qcompressive strengthGonzatti C a, Zorzi L., Agostini I.M. Fiorentini J.A., Viero A.P. Philipp r.P. aartment of Mineralogy and Petrography of the Federal University of Rio Grande do Sul, BrazilFoundation for Science and Technology of Rio Grande do Sul State, BrazilARTICLE INFOA BSTRACTArticle historIn the early 1990s, the Foundation for Science and Technology of Rio grandeState(CIENTEC)Received 18 January 2014developed a pioneering study in Brazil, related to the simultaneous mining of multiple coal seamReceived in revised form 10 March 2014Accepted 22 MayOne of the activities included detailed studies on the geomechanical characterization of materials presentAvailable online 14 November 2014in the Irapua coal seam, under exploitation in the A-Sangao Mine, located near the city of Criciuma-SCthe uniaxial compressive strength of the Irapua coal seam and establish a first approximation for theSize effectin situ strength value of this coal seam, since existing knowledge is solely based on practical miningexperience over the years. Large samples of the coal seam were collected using special techniques toCoal seammaintain the integrity of the material, and a set of 56 uniaxial compression tests in cubic specimens, withside length ranging from 4.5 to 31 cm, were conducted in laboratory. This paper describes the experimentechniques used in the assays, and also presents the uniaxial compression strength results obtaineMoreover, important aspects of this type of study are considered, highlighting the size effect for thecarbonaceous bed and the estimation of in situ strength values for the Irapua coal seam.o 2014 Published by elsevier B V on behalf of China University of Mining TechnologyIntroductionIn summary, the size effect implies higher strength values forsamples tested in laboratory, when compared toAlmost all of the scientific production on what is still consid- This increase in strength is attributed to the lower presence ofred the art of scaling pillars in coal mines, originated in the U.S. surfaces of weakness in the samples used in laboratory tests, sinceand South Africa countries, where underground coal mining, by the small dimensions used do not entirely represent the mass thatthe chamber and pillar method, surpassed 90% of the total amouroriginated them.produced until the early 1990s.In the case of carbonaceous rocks the definition of the charac-Considering the two necessary parameters for the analysis of teristic strength of the coal seam is of paramount importance, sincese safety level of a pillar-strength and load-the first is the one that it is the very rock that composes the support structures(pillars)ofas deserved the most attention from many researchers who have the excavations from where the ore is extracted. In this type ofdedicated themselves to this subject, and the latter for practical rock, this is a complex task, since the size effect seems to be evenpurposes, is basically restricted to the application of tributary area more pronounced that in other materials, due to the formationtheorycharacteristics of the coal, as reflected by the presence of smalPillar resistance, defined by the maximum load supported by to large geological structures, such as cleats, bedding planesunit of area, depends on at least three basic elementsfractures, faults and inclusions of material other than coaDifferent techniques for the estimation the in situ strength ofThe size effectthe coal seam can be used. The technique of laboratory tests withThe shape effect;small samples is among the less costly ones from the economicalne properties of the rock that forms the pillar and the host point of view however, it is limited due to the low representationof the lithological variations and structural characteristics of therock mass. When there is access to the coal seam, using largersamples than theinre ic more desirrable andCorresponding author Tel +55 51 98939643tends to substantia中国煤化工 of the mechanicalE-mail address: gonzatti@ufrgs. br (C. Gonzatti)CNMHGhttp://dx.doiorg/10.1016/j.ijmst.2014.10.0032095-2686 2014 Published by Elsevier B V on behalf of China University of Mining TechnologyC Gonzatti et al /International Journal of Mining Science and Technology 24(2014)747-754behavior of the rock in order to estimate the in situ resistance with addition to serving as a base for the proposal of a methodologyeasonablefor pillar sizing applied to southern Brazil coal seam conditionshe use of laboratory testing technique, with variablee 9]. Once the technique of laboratory tests was defined as the pathsizes, has been done by many researchers worldwide for over chosen for estimating the in situ strength, the representativity of100 years of research Studies on mineral coal have been conducted the samples regarding a particular rock mass is an important factorsince approximately 1875 [1. Steart, based on laboratory testing in for the application of the results in the sizing of the undergroundsmall samples and underground observations, and Gaddy, structures. Generally, the ideal size(theoretical) runs into theindividually studying several layers of American coal, concluded limited operational capacity of the equipment available forthat the strength behavior is inversely proportional to the size of extraction as well as the performance of the laboratory teststhe sample [2, 3 They suggested the following generic function The following topics show the results from studies conductedto explain this behaviorfor the estimation of the in situ strength of the Irapua coal seamof the a-Sangao Mine, located in the South-Catarinense coalfieldin southern Brazil, during the development of MULTICAMADASproject [10 The work performed to prepare the specimens,where oc= uniaxial compressive strength of a cubic test specimenequipments and the results are described. Finally, an estimatik= constant dependent on the physical characteristics of each coal of the in situ strength of Irapua coal seam is suggested to use inseam;D=size of the tested cubic test specimen; a= coefficient the analytical models to pillar dimensioningcorresponding to the size effect for the studied coal seam, with2. Geology of study areavalues of 0.5 representing the mean for the different studied layersRegarding British coals, Evans Pomeroy conducted a largeGeographically, A-Sangao Mine is located within the Southamount of uniaxial compression tests, using cubic test specimenCatarinense coalfield, which is part of the Eastern Edge of the Par-with base sides ranging from 0. 3 to 5.4 cm and prismatic ones, withbase sides ranging from 1.3 and 5 cm and height between 1.3 and ana Basin, as seen in Fig. 1. This basin, which in Brazil s gentry5.4 cm [4]. Bieniawski in turn, reports an extensive stuseveral other states, is of Devonian age for its oldest sedimeInd has the evolution characteristics ofconducted in South Africa, with samples from the Witbank coal basin. Within the stratigraphic sequence, the carbonaceous litholseam 5]. This author was one of the first to study the size effectogies are linked to the units called Itarare Group(more basal)to uniaxial compressive strength in coal based on tests on cubic and Guata, with ages between the Upper Carboniferous and Uppertest specimens, with sizes comprised within a wide range-from 2 Permian. But it is in the Guata unit that the thicker coal seams areto 2 m-and all tests were conducted with underground infrastrulocated of better quality and under exploitation. among the seventure. For him, the decrease in strength as a function of the increase individual coal seams in the State of Santa Catarina, three haveof the sizes of the test specimens is associated to the greater prob- more expressive economical importance: Barro Branco, Irapuaability of more discontinuities in larger test specimens than in and Bonito seamssmaller ones. The approach by Protodiakonov Koifman suggestssome explanations for the decrease in rock strength with the is represented by normal faults, with subvertical planes. Thrustvolume effect" and the surface effect'faults rarely occur, with little tailing. The presence of"slickensidescaused by differential sediment compaction is frequent. FractureIn Brazil, the first studies with coal using scientific techniquessystems can follow faults, forming conjugated systems, filled orcome from the early 1980s, with emphasis on the laboratory and not by diabase dikes. The folds seen in some coal seams are assoin situ work developed by the Institute of Technological Research ciated to"drag fold"faults. However, folds caused by compressiveby the foundation for Science and Technology of Rio grande doStratigraphically, the Irapua layer is below the barro brancoSul State( Cientec)[7, 8. Cientec studies, in about 1300 telayer. The thickness of the cover ranges between 95 and 145 mspecimens, covered the most important coal seams under undeincreasing to thicker coverage as the layer approaches the flowsground exploration at the time or with future potential. This worfrom the Serra Geral Formation. The sterile interlayer intervalwas pioneering for the definition of in situ strength of the I F layers ranges between 12 and 16 m, with an average thickness of 13 m(Charqueadas deposits, in Rio Grande do Sul state)and Barro Fig. 2 shows a typical profile of the Rio Bonito Formation and theBranco (South-Catarinense deposit, in Santa Catarina state), in Irapua coal seam in the study area.The Irapua layer is of restricted occurrence and appears on plantview in the shape of a string, composed by isolated segments, withIn profile, itshape where the layer tapers towards one of the flanks. In someplaces of the mine, this shrell defined by changes in the dip-ping of the floor or roof planes of the coaeamreaching10°ofdipping, with thickness variations of more than 1 m. Its contactswith the floor and roof sandstones are sharp, forming well definedpartition planes. This trough shape is associated to the possible tuffdeposition in paleochannels. The mining area where this study was图 Quaternary sedimentsdeveloped is located in one of these paleochannel segments. Theaverage thickness in the mined section is 1.77 m, reachingGonduanic sediments2. 60 m in some loE- Devonian sediments中国煤化工感 Shield pre-devoni3. Sampling and preCNMHGnensFig. 1. Geological map of South Brazil(Numbers from I to XIlI indicate knownThe samples were collected during the mining of the Irapuacarboniferous Basin in Brazil)[11-seam With the aid of a universal cutter, blocks of up to 0.4 mofC. Gonzatti et al/ International Journal of Mining Science and Technology 24(2014)747-754South catarinense coal basin profileNe IRAPUA coal seam profilesiltstoneDull to bright coal with thin bladesTreviso coal seamvitrenia pyrite nodulesIRAPUA coal seam6Ponte Alta coal seamblades vitrenio pyrite nodules“C” coal seanFig. 2. Typical stratigraphy profile of the rio Bonito Formation(a)and lrapua coal seam at A-Sangao Mine(b).volume (approximately 0.7 m x 0.7 m x 0.7 m) were separatedfrom the mass by successive vertical and horizontal cutsImmediately after extraction, the samples were covered with athin layer of wax to protect the rock against the variation ofhumidity and the action of weathering agents. the blocks weretransported on sawdust mattresses in order to minimize the damage by shock during the journey from the mine to the laboratoryThe preparation of cubic specimens with different sizes wasdone withe of large diamond saws, as shown in Fig 3, andaller saws were used for the final cut and finishing of the testspecimens. Fig. 4 shows test spein different size rangebefore failure. In total, 56 cubic test specimens were prepared, withL= 20cmL=25cmthe side dimension (L)ranging between 4.6 and 31. 4 cm, in theamounts seen in Table 1Fig. 4. Cubic test specimens in larger size ranges used in the study of the size effectL是Qon the Irapua coal seam strength[101In order to provide additional information to help definingsize effectal strength bulkand p-waI bedding pladetermined, using specific methods for each test, recommendedby the ISRM [12-14. Determination of ultrasonic velocity wasdone with the use of MkV PUNDIT equipment, with 1 MHz,200 kHz and 54 kHz transducersted in a hydraulicpI中国煤化工eand a reaction gantCNMHly, with load cap.test specimens.Fig. 3. Saw with 1.0 m diameter blade, used for reducing the size of the coalUniaxial compressive strength(UCS) was determined by apply-ing the load on the test specimen in the perpendicular direction toC Gonzatti et aL /nal Journal of Mining Science and Technology 24(2014)747-754Table 1Table 3Sizes and amounts of cubic test speprepared for assay.Summary of results of the P-wave velocity by cubic test specimens size range.L(cmverage sizeStandard deviatiL(cmNVp(km/s)CV(‰)of cube side(cmthe side ( cmMean2.340.2614.7-15.60.3136-20.620.10.2124.6-25.6250.71.5431,1-3120.21.58sImensTable 4Summary of UCS results by cubic test specimen,s size rangSummary of bulk density results by cubic test specimeL(cmN(‰n sIze range.MeanCV(%10.0165215.13.391.6220.12681.5725.110.5931.394631.3Note: SD- standard deviation: CV coefficient of variationThe variability of the relationship between uniaxial compressive strength and P-wave velocity does not depend on the testThe variability of the strength parameter, between 21% and 27%the bedding planes of the coal seam, according to ISRM recommenis higher than that of ultrasonic velocity(form 12% to 16%)dations for this type of assay[13]. Typically, the loading speed usedresulted in a stress rate in the order of 0.7 MPa/s4.2. Correlations with regression analyzes3.2. ResultsTable 5 shows the linear correlation between the variablesuniaxial compressive strength, P-wave velocity, bulk density anda summary of results related to the physical indices(bulk den- side dimensions of the cubic test specimens. The correlationssity, p, and P-wave velocity, Vp)can be seen in Tables 2 and 3. As to between the variables ocxL, dc x Vp and Vp x L are significanthe uniaxial compressive strength, o, they are shown in Table 4ibility of thThe simple regression study between the pairs of variables thatshowed significant correlation resulted in the following equatand strength, separated by test specimen size range, in terms ofnedian and quartiles4. Analysis of results and estimate of the in situ strength of theed variables, with correlation coefficient. amount oftests in brackets, and significance levelIrapua coalseamG(MPa)L(cm)p(km s4.1. variabilityAs to variability the observation of variation coefficients, which0.00best summarize variability, in Tables 2-4, showsVp(km/s)(54)to composition, the material tested is relatively homogenp(g/cm)0.620.16since the variation coefficient of the bulk density is between 3%6(54)and8‰0.001.9615T1,5Outliers1015,120.125.131.35.51015.120.125.1313Cube side dimension(cm)中国煤化工(a)Effect of test specimen size onb)Effect of test specimen size on the P-waveCNMHGn thevelocity perpendicular to thebedding planesFig. 5. Variability of the physical indices and strength, separated by test specimen size range, in terms of median and quartilesGonzatti et al International Journal of Mining Science and Technology 24(2014)747-754GMPa)=3786L(cm)038(1in the statistical analysis of the adequacy of a regression model arshown in the graphic representation of the modelsVp(km/s)=2984-0.423lnL(cmAlthough the coefficients of determination, r show that theG(MPa)=887p(km/s9084(3) variability displayed by the models is lower than 50%, in all cases,In Figs. 6-9 are plotted the observed data, the regression model coefficient of correlation and null hypothesis of the coefficientscurves, and the confidence interval of the mean and of an individual at significance level of 0.05%, indicate that the models are usefulvalue for the confidence limit of 95%. The main parameters involved to explain the data30a201020Cube Side Dimension(cmParameter CoefficientStandardSignificance level0.377-6.188684E-8Correlation coefficient=-0 64Coefficient of determination=41.5OriginfreedonSignificance leveldel54Fig. 6. Size effect of the uniaxial compressive strength.3.228兰91.2Natural Log of Cube Side Dimension(Model: V=a+b*In(l)Parameter CoefficientdeviationSignificance level0.178516.71070.42260.07324.4ECorrelation Coefficient=-0 62Correlation of Determination=39.0%Parameter CoefficientStandard中国煤化工deviationT value298241785CNMHG00732彐YHFig. 7. Scale effect of the P-wave velocityC Gonzatti et aL /International Journal of Mining Science and Technology 24(2014)747-754E兰3.乏360202030Model: O VIParameterCoefficienttandardvalueSignificance level120.660.1910-0.584974605Correlation Coefficient=-072Coefficient of Determination=51.7%0Analysis of VarianceOriginSum offreedomSignificance level4.9932Fig 8 Size effect of the uniaxial compressive strength multiplied by the P-wave velocity1.21.62,42.83,2Model: o =aVpParameter CoefficientStandarddeviationT valueSignificance level020694.03271.8E-4Correlation Coefficient=0. 49Coefficient of Determinanalysis of varianceSum ofOriginSignificance leveluarefreedomModel1.287316.260001841161Fig 9. Regression between uniaxial compressive strength and the p-wave velocityIt is observed in Figs. 6 and 7 that as the test specimen size regression is made as a function of the cube dimensions. Figincreases, the compressive strength and ultrasonic velocity shows the regression curve and statistical analyzes. There is andecrease. The size effect that manifests due to imperfections in increase in the coeff n中国煤化工64t0072.Thethe test specimen structure affects both compressive strength model then explainsand ultrasonic velocity, as expectedOn the other haCN MH Gd between theThis mutual effect is more intense when the observed strength uniaxial strength and p-wave velocity is not very expressive, aand velocity values are multiplied among themselves and a shown in Fig 9, with a correlation coefficient less than 0.5C. Gonzatti et al/ International Journal of Mining Science and Technology 24 (2014)747-754Table 6x "exponent values in the size effect function obtained for different coal seamsCoal seamx" value in function oc∝吉 Referencesa=3786*L038Llanderbie(England)0.4415G=50.27*L05Betteshanger(England)047B:::Langwith(England)0.53linwood (USA)Pocahontas n。4(USA)Cube side dimension (cm)Marker (USAFig. 10. Variation of size effect considering different models for Irapua coal seam.[16Deep Duffryn(England)Barnsley Hards(England)Witbank(South Africa)[183530llF coal seamNote:# Theoretical strength based on the concept of energy balance, derived forIn situ test results- Barro Branco coal seam154.3. Model proposition5As initially stated, the size effect in the strength of coal testCube side dimension(cm)specimens has been studied since the late 19th century by a greatnumber of researchers in different countries, each one of themFig. 11. Comparison of the size effect of the Irapua seam with other Brazilian coalinvestigating coals whose characteristics are not always similaThese multiple authors sought, through laboratory assays withcubic test specimens of different sizes, to define with greater clar-ity the law that explained the behavior of the coal strength as a Table 8function of the increase in sample size, that is, the size effect in In situ compressiveestimated for the lrapua coal seam considering differentthe various coal seams. Table 6 summarizes the results obtainedfrom a large amount of research conducted by different authorsModelacm at critical size(MPa)on carbonaceous materials, which agree with the law proposedLari=90 cmLerit =150 cmby Steart and Gaddy [2,3]a)ocm=37.86*L0385.64On the other hand the behavior envisaged by these authors and(b)oem=50.27*1-05many others over time, does not occur indefinitely that is, there is(c) According to Wilson [2214.15sample size from which the strength becomes constant or variesvery little. This limit, called"critical size", is the size from which,for practical purposes, the coal strength remains constant. Deter- producedranging from 0.06, virtually inexistent, to 0.23mining this size was one of the greatest advances for the project. The use of a=0.5 was opted for more conservative results. AnSeveral suggestions have been proposed by different researchers, exceptionBarro branco seam. whose recommended a difs can be seen in table 7ferent modelRegression study with experimental data in the Irapua seamFig. 11 shows the curves adopted for the barro branco(in Santashowed values of k=37.9 and a=0.38, as shown in Eq (1). When Catarina state)and I F(in Rio Grande do Sul state) seams, thefixing the a value in 0.5, k=50.27 is obtained. Both curves are results of in situ tests executed in large blocks of these seamsshown in Fig. 10. The vertical dotted lines, traced from the x-axis, and the Irapua seam curve, with a=0.5indicate two of the most recommended critical sizes. StartingThe method by Wilson adopts as the in situ strength of the rock,from the first size(90 cm) to the second(150 cm), a strength the strength value determined in laboratory divided by a reductiondecrease in 23% is seen, in the case of a=0.5, and of 17%, when factor that varies as a function of the lithological type and fracturx=0.38,degree present in the rock mass [25. In the case of the coal, heIn research conducted by CientEC to establish the method of recommends a reduction factor of 5pillar dimensioning, assays were conducted in four coal seamsThe different estimates for the in situ strength of the IrapuaBarro Branco, I, Chico Loma and l[9]. The size effect investigation seam are summarized in Table 8. For this coal seam, equation b)Table 7Suggested critical size values for the estimate of in situ strength of coal seamCritical size for the estimate of in situData base for propositionReferencesLaboratory and in situ tests in test specimens with sides rangingaboratory tests with cylindrical test specimens with diameterH中国煤化工ust1fCNMHGaboratory and in situ tests in test specimens with sides rangin0.9to1.0754C Gonzatti et aL /International Journal of Mining Science and Technology 24(2014)747-754from Table 8 is suggested, with a critical size, Lcrit, of 90 cm, result[5 Bieniawski ZT. The effect of specimen size on compressive strength of coal. Int Jing in a value of 5. 30 MPa for the in situ strength of a cubic pillarRock Mech min Sci 1968: 5: 325-35of rock andcoal S Afr Counc Sci Ind Res 1967: 3: 97-1085 Conclusions[7] Midea NF, Yoshikawa NK, Marchi A]. Pillar design-a example of the role of rockchanics. In: The 1st Brazilian Mining Congress, Brasilia, Brazil, 1985. pp198-14The studies conducted in laboratory for the estimate of the [8) Zorzi L. PILAR projet: pillar design in coal mining. CIENTEC-Foundatiin situ strength of the Irapua coal seam, of the South-CatarinenseScience and Technology of the rio grande do sul State 1990: 2: 518Carboniferous basin in southern brazil showed several difficulties[9 Zorzi L, Agostini IM, Gonzatti C. Methodology for pillar sizing applieuthern Brazil coal seam conditions. CIENTEC-Foundation for Scienin determining a representative value for the coal seam.Technology of the Rio Grande do sul State 1991: 23: 31he technique based on laboratory tests on cubic test specimens [10 Zorzi L MULTICAMADAS Projet: underground mining of multiple coal seams-of different dimensions proved very useful, despite the greatgeomechanical study. CIENTEC-Foundation for Science and Technology of thdifficulty in extracting, maintaining the quality and preparing the (11] Costa JFCL Evaluation of coal reserves-Fontanella Projet-Carboniferasamples, a need for compatible sample sizes with the equipmentMetropolitana S.A. Mining Engineering Department, Federal University of Rioavailable in the laboratory result variability and the impossibilityof representing the different geological-structural patterns existing112 ISRM. Suggested methods for determining sound velocity, Int Rock Mech MinSci Geomech Abstr 1978: 2 (15 ) 53-8in the coal seam[13 ISRM. Suggested methods for determining the uniaxial compressive strengthStill. the results obtained allowed to estimate. for the areand deformability of rock materials. Int J Rock Mech Min Sci Geomech Abstr979:2(16):135-40corea Th sestig taes urt kmsistenm ith s e es z s a. v an[14] ISRM. Suggested methods for determining water content, poabsorption and related properties and swelling ansuggested by Zorzi, obtained by previous analysis of pillar failureproperties. Int Rock Mech Min Sci Geomech Abstr 1979: 2(16): 141-56.cases at the a-Sangao Mine itself [9[15 Millard D), Newman PC, Phillips JW. The apparent strength of extensivelyAlthough the technique of testing samples of different dimeEngland.1955;(68):723-728sions is labor intensive and seems to be currently out of use, it is [16 Skelly wA, Wolgamott J. Wang FD Coal mine pillar strength and deformationto dismiss it as an alternative to estimate thprediction through laboratory sample testing. In: 18th Proceedings of the U.SSymposium on Rock Mechanics, Keystone, Colorado, USA, 1977: 2B5-1-2B5-5in situ strength, especially in the initial stage of the operation of [17] Evans 1, Pomeroy CD, Berenbaum R The compressive strength of coal Collierya new mine in coal seams that are mechanical and structurally [18] Bieniawski ZT, Van Heerden WL The significlittle knownf in situ tests on large rockRegardless of the limitations imposed by the technique of (19) Fairhurst C Laboratory testing of rock and its relevance to mine design. SMEtesting the materials in the laboratory the benefits from the uselin Eng Handb 1973: 1: 13-36of this device are undeniable, even if the laboratory technique is20 Hustrulid WA. A review of coal pillar strength formulas. Rock Mechnot the only one, nor the ultimate way to obtain information about (211 Barron K, Yang T Influence ofsize and shape on strength of coal. Partthe mechanical properties of a rock massuniaxialof coal. In: Workshop on Coal pillaReferencesPittsburgh, PA, USA. 1992. p 5-24.22 Bieniawskiconditions. In: International Conference on Stability in Undergroundt, vancouver, BC, Canada. Society of Mining Engineers of AlME. New York,Mechanics, Rolla, MO, USA, 1964, p. 450-6USA.1982.p.1949.[21 Steart FA. Strength and stability of pillars in coal mines. J Metall Min Soc S Afr [23] Bieniawski ZT. New design approach for room-and-pillar coal mines in the1954;3:30925ISA. In: International Congress on rock5th. Melbourne australia3 Gaddy FL a study of the ultimate strength of coal as relaDsolute1983p.E27-E3ze of the cubical samples tested. Virginia: Virginia Polytechnic Institute: [24] Bieniawskiin mining andtunnelling. Rotterdam: Balkema: 1984. p. 55-96.[4] Evans L, Pomeroy CD. The strength of cubes of coal in uniaxial compression. In: [25] Wilson AH. The stability of underground workings in the soft rocks of the coalof non-metallic and brittle mateleasures. Int Min Eng 1983: 2(1): 107-8Butterworth Scientific Publications, 1958, p 5-28中国煤化工CNMHG