Determination of coal mine mechanization using fuzzy logic

bleonlineatwww.sciencedirectconMININGScienceDirectSCIENCE ANDTECHNOLOGYining Science and Technology 19(2009)0149-0154Determination of coal mine mechanization using fuzzy logicAtaei M, Khalokakaei R, Hossieni MFaculty of Mining, Petroleum and Geophysics Engineering, Shahrood University of Technology, Semnan 3619995161, iranAbstract: Two of the most important tasks in coal mines are to improve efficiency and to increase production besides keepingsafety constantly in mind. In order to obtain these goals, mine mechanization is required. Mine mechanization needs high levels ofnvestment and should therefore be studied carefully before final decisions about mechanization are made. When analysizing thepotential for mechanization the following, rather imprecise, factors should be considered: seam inclination and thickness, geological disturbances, seam floor conditions, roof conditions, water at the working face and the extension of seams. In our study wehave used fuzzy logic, membership functions and created fuzzy rule- based methods and to considered the ultimate objectivemechanization of mining. As a case study, the mechanization of the Takht coal seams in Iran was investigated. The results show ahigh potential for mechanization in most of the Takht coal seams.Keywords: fuzzy set; mechanization; coal seams: productivity; safety1 Introductiontions: these may be split broadly into longwall andpillar methods of workingThe fundamental task of mining engineers is toThe longwall coal mining method is widely perproduce more coal at a given level of labour input ceived to be a twentieth century development; howand material costs, optimum quality and maximum ever, the concept is actually believed to have beenefficiency. To achieve these goals, it is necessary to developed more than 200 years ago. Its main advanceautomate and mechanize mining operations. Mecha- is that earlier operations were, in principle, manual,nization is an objective that can result in significant while since the 1950s, the level of mechanization hascost reductions and higher levels of profitability for increased to the stage that today longwall faces areunderground mines. For these reasons mining engi- highly productive units which can be operated by aneers are continuously looking for different ways to very small work crew. This method has a simple sys-mechanize mining procedures, especially the mecha- tem layout and provides for continuous productionnization of underground coal mines that provide a and full potential for mechanization, which can fur-large potential for reduced costs, increased safety and ther improve productivity, personnel health andimproved profitability. Mechanization and automa- safety. Mechanized longwall rather than other coaltion are changing the way in which ore deposits are mining methods are more expensive to equip andbeing mined. The industry needs more energy-effi- operate, but are in general more productivecient methods, systems and approachesIn this study we have reviewed some of the factorsFor many years, machines have been used to ex- affecting the development of mine mechanization,ascavate soft minerals in underground mines on a con- well as some other factors that will be required fortinuous basis. In some cases, this includes soft or successful mechanization in the future. As a caseweak rocks, surrounding the valuable minerals which study, we present the mechanization of the Takht coalare being extracted. Coal is probably the most com- seams in Iranmon exarThe methods of working coal seams have gradu- 2 Objectives of mechanizationprogressively improvemodified as knowledge and experience are gainedThe most important objectives of mechanizationand power machines become available. Over the are followced costs; faster development; fasteryears,a very large number of coal mining methods mini中国煤化工 ng production athave been developed to suit seams and local condi- fewerproduction ratesCNMHGReceived 04 May 2008: accepted 16 September 2008Corresponding author. Tel: +98-273-3335509: E-mail address: ataei@shahroodutac irMining Science and Technologyvol 19 No. 2per shift; mining with smaller underground crews; placement index (I), a parameter, to quantify geomaller capital expenditure per extracted ton of coallogical disturbances as followswork under protectively supported roofs; develop-ment of more productive crews3 Factors affecting the development of and m is the thickness of the seam(m). table I cm)where t is the displacement of a seam by faultsmechanizationthe level of seam uniformity with respect to the indexof displacement. In this classification, seam uniform-In order to take advantage of mine mechanization, ity ranges from 0 to 1, where seams with an indexeither in an entire mine or at least a section of a mine, 1, =0 are completely uniform and seams with a dis-some factors must be considered. Seam gradient andthickness, geological disturbances, seam floor condiplacement index of more than 3, are considered to benon-uniformtions and roof formation are some of the most im-portant factors that affect the potential of coal seamsTable 1 Proposed, classification for determinationSeam uniformity3.1 Seam inclinationUniform06-1.000.5Most longwall mining activities occur in flat oreSemi uniform0.2-0605-15deposits or nearly flat seams. In general, seams with00.215-3.0low a dip(zero or near-zero dip angles) are moreamenable for mechanization With increased sear3. 4 Roof conditionsgradients, the application of mechanization becomesmore difficult. Although the best operational condi-Roofs should cave. If a roof does not cave ortions are on level seams, in rare cases seams with up hangs, it is not suitable for longwall miningto 80 degrees of inclination have been worked, alit may fall unexpectedly. Thus, some stowing systemthough more usually seams up to 35 degree of inclishould be used. The most suitable roofs cave in as thenation can be mechanized by power support.support advances. However, when a very weak roofcrumbles rather than holds, part of the coal seam is2 Seam thicknessleft to support it.The thickness of the seam and its regularity is anBoth operational experience and research resultsimportant determinant of the level of mechanization have demonstrated that roof stability is relative. Forthat could be applied at a typical coal face. Supports an unstable roof, certain techniques are required tocan be lengthened and further increased by some ad- control and change the factors contributing to theditions.However,great irregularities cannot be ac- unstable conditions and to upgrade its stability. For acommodated. Cutting low sections may be difficult medium stable roof, a power support is required,for the winning machines and washing facilitiesproperly selected and applied, to maintain roof stabilTherefore, a good survey of the seam should be made. ity. Often a medium stable roof becomes unstable duebefore choosing proper support. The thickness that to periodic weighing, poor roof conditions or im-an be worked, at present, ranges from 0.6 to 5 m properly selected power supports. For stable roofs,with most seams between I and 3 m. In seams thicker with the exception of a gradually sagging roof, theirthan 3 m or thinner than 0.6 m, the possibility of stability must be destroyed artificially and systemati-mechanization will be reduced. Thicker seams should cally in order to avoid large aerial caving.be worked out in slices or by recovering the cavedOther quantitative methods are available for evalcoal from the back by special arrangementsating the propensity of roofs to cave in. These meth3.3 Geological disturbancesods employ various factors such as lithological sequences, amount of roof convergence at the gobThe effect of faulting on the geomechanics of edge, lack of support over a certain time period belongwall is one of the most difficult issues to predict. fore caving, seismic wave velocity, drill coreIn some cases, the presence of faults or jointing can strength, average frequency of bedding plane andhave a dominant effect on the geomechanics of a re- rock strength and bed separation resistancetreating longwall. If there are complex geologicalUnrug and Szwilski (1982)made recommenda-conditions suchas faults and seam pinch-outs, the tions concerning the principal strata control paramepossibility of mechanization will be reduced. Theters for the design of longwall faces. Their empiricalamount of coal seam displacement and the number of for中国煤化工 strength index isfaults present over the length of a seam are very im- giverportant factors that effect the condition of the work-CNMHGing face and the decision to mechanize the operation=U.010×CX1X2×3of the seam. In our study, we have defined the dis- where e, is the roof strength index, 0 the averagetaci M et alDetermination of coal mine mechanization using fuzzy logiuniaxial compressive strength of the core(kg/cm2), immediate roof (cm)and K a swelling coefficientK a factor to account for a decrease in strength from with a value between 1.3-1. 5. The various designthe laboratory to a field specimen, K, a factor to parameters are based on a roof classification systemaccount for a decrease in strength with creep load- represented by the roof strength index(see Table 2)ing, K, a factor to account for decrease in strength Table 3 shows the values of different factors forwith an increase in humidity, m the thickness of the various types of roofsTable 2 Caveability classification based on roof strength and time exposureRoof typeRoof strength inder0≤s18After exposure, roof caves in immediately or after a short delayLow stableRoof very difficult to control. Full of cavities, fractures and fissures, caves in easilyMedium stable35≤Qs60Caves in easily. From fractured roof with local falls to fairly good rox605Qs130Good roof with excellent caving properties to hardly any cavingVery stableQery strong and very stable, Artificial caving is necessaryTable 3 Values of different factors for different rocksrocks supports. It takes 15-20 days to install the equip-Rock typement,which adds to the cost of the coal; the largerthe panel the lower the installation cost per ton ofK,extracted coal. However, very large panels may beexpensive in upkeep of the gateways, although this isnot an insurmountable problem. The optimum width3.5 Floor conditionsis found to be 800-1000 mThe floor should be strong enough to resist intru-sions Intrusion of soft floors is troublesome for ad. 4 Use of fuzzy logic for estimation of me.ancing and also makes the roof conditions difficulthanization potentialto control owing to high convergence. Some coalWhen evaluating the effect of each factor on pomay be left if the coal is hard.tential mechanization, the importance of each ofThe reaction of floors to any kind of support, in- these factors should be defined. This definition isstalled along or behind longwall faces, significantly subjective. When technically assessing the result withaffects strata stability. If the design of the support is respect to geological factors, qualitative terms areto be based on an acceptable rate of closure or de- commonly used, e.g. excellent, favorable, poor, etcformation along a longwall face and its ends, then, in These descriptive terms are ambiguous and impre-order to ensure support balance and stability, the stra- cise The question that is raised immediately is actum pressure within the face region should be con- cording to what criterion(or scale); does'excellent'trolled. This requires: (a)uniform pressure and de- or'favorable'apply?formation distribution along the face; (b)a floorFuzzy logic is a powerful tool for the analysis ofbearing capacity in excess of the effective stratum systems that work with vague parameters and whichpressure exerted upon it through the supports 2.receive qualitative inputs, uncertain and simple ana-Where footwall rocks are weak, support systems lytical information of the conditional IF-THEN typemay fail by punching into the peripheral rock of ore in an algorithm in the shortest time and with the leastbodies. The failure mode is analogous to bearing ca- calculated success of suitable results. Hence in ourpacity failure of a foundation and may be analyzed as investigation we tried to present a way for the deter-such. The floor rock bearing capacity is direct related mination of potential mechanization of coal seamswith the uniaxial compressive strength of rocks. In with fuzzy logicgeneral, a higher strength ima greater bearingcapacity and a greater potential for mechanization of 4.1 Basic description of fuzzy setsthe seamSince the notion of fuzzy sets was first introduced3.6 Water at the working facein 1965 by Zadeh, the theory of fuzzy sets has accelerated considerably. Its potential usage is being con-Water at the working face is detrimental and corro- stantly explored and documented". The applicationsive to the supports. Under these conditions either the of this approach to subjective decision-making in thepanel should be drained by drilling, or special anti- mining industry is slowly being investigated. The keycorrosive supports should be chosen. Water is always concea handicap and miners hate to work under wet condi- of中国煤化工 y is the definitionmapping of a ref-ereneCNMHGu determines a3.7 Extension of seamfuzzy subset, denoted by A, in U. A is called theThe panel should be large enough to use powered membership function, sometimes indicated by PAMining Science and TechnologyVol 19 No. 2uA(u)is called the grade of membership of u in A.If u,(u)is either 0 or 1, the fuzzy set a will beLflOwderstood as an ordinary subset4.2 Membership degree as effective parameter inpotential coal seam mechanizationSeam inclination is one of the major factors0.25which ultimately determine the potential of mecha-Seam uniformity scorenization of coal seams. A fuzzy membership grade ofseam inclination has been developed, as shown in FigFig 3 Membership function diagram for geologicalIt is to be noted that"very low"is allocatedembership grade of 1.0 at a seam inclination s 8degree after which it gradually declines to 0. On theother hand, "very high"means a membersiof 0 for a seam inclination <45 degree andincreases to 1.0 at a seam inclination >50如四时06ery highOther qualitative legends(low, medium and high)aredefined as shown in Fig. 1.8.0High Very highFig. 4 Membership function diagram for roof conditions0.008Seam inclination(degreeg. 1 Membership function diagram for seam inclinationSeam thickness is defined between 0 and 10 m Figes five qualitative legends "very lowl1001500low”,“ medium”,"high”and" very high”."LowBearing capacity(MPa)is allocated a membership grade of 1.0 when thethickness of the seam is between 0.8 and I m and 0.0ig. 5 Membership function diagram for floor conditionswhen the seam thickness is≥14mor≤04m. Otherualitative legends are defined as shown in Fig. 2.For geological disturbances three qualitative legends(low, medium and high) are defined as shown in Fig3. For seam roof formation and floor conditionsqualitative legends are defined as shown in Figs. 48642and 5. For water at the face also three qualitativelegends (low, medium and high) are defined asshown in Fig. 6. For seam extensions qualitativelegends are defined as shown in Fig. 7. Fig 8 showsWater inflow(m per minute for each 10 m of face length)seam membership grades of potential mechanizationFig. 6 Membership function diagram for water atdefined by five qualitative legends(very low, lowmedium, high, very high)864中国煤化工CNMH500020000m thickness(m)Fig 2 Membership function diagram for seam thicknessFig 7 Membership function diagram for extension of seamAtaei Met alfuzzy variable, called ore quality (Table 4). The sec-ond rulle-base includes 75 rules for its inputs, i.e.,floor quality, roof quality and water at the face. Itsoutput is a fuzzy variable, called rock quality (Table5). Finally, the third rule-base makes for 25 ruleswhose inputs are ore quality and rock quality andoutput is the potential mechanization of a seam(table 6).Fig 8 Membership function diagram for potentialmechanizationTable 4 Part of first rule-base4.3 Fuzzy rule-baseFor estimating the potential for mechanization,based on available inputs, a number of simple rules Very low Low Very low=ery lowin the form of IF-then statements are needed torelate inputs to suitable results. In this study threevery lotrule bases are shown in Fig. 9. Therefore our firstrule base consists of 225 rules concerning its inputs,i.e., seam inclination, seam thickness, extension ofseam and geological disturbances. The output is aUniformityMAXXAUnifRock quality MechanizabilityMecharizabilityMAXXL Roof qualityRoof quality Rock qualityAL WaterWaterMAXFig 9 Rule bases for potential mechanizationTable 5 Part of second and third rule-baseSecondTHENTHENRoof qualityRockqualityOre qualityRock qualitylechanizabilityLow4. 4 Fuzzy inferenceCOM methodThere are many ways of inference. For examplemethods of inference, when the if part is invoked, 5 Case studyare GAMMA, MIN-AVG MIN-MAX and methodsof inference of the then part are MAX, BSUM. InThe Eastern Alborz coal basin is locaour case, we opted for the MAX-MIN operatorerm Iran and presently includes all the carboniferoussediments contained within the bounderieas formed4.5 Defuzzificationby the Firoozkooh road and the Bojnord-EsfarayenAfter the fuzzy output has been determined, a cer- road中国煤化工include the car-tain level of defuzzification of the potential of boninof the centralmechanization must be carried out. At this stage, a deserCN MH Grin-Jam and Lar-choice of different methods, such as the center of estan and on the east side our study area continued tomaximum(COM)and the mean of maximum(MOM) the Sarakhs. The Takht coal mine is one of a numberare available: for our study, we have opted for the of mines in the Eastern Alborz coal basin, situated tothe south east of Minoodasht city in Golestan Prov- the potential for mechanization of the Takht coalince (Iran)seams. Table 6 shows the result of our study of theBy using the fuzzy Tech 5.12 software, we studied potential for mechanization of these coal seamsTable 6 Potential for Mechanization of the Takht coal seamsSeam Seam inclinationFloorUniformity qualiqualityinflowextension32.51404112.56416410.580p01834625354261595692305508525Very low6 ConclusionsReferencesCoal has always been considered an important [1] Unrug K, Szwilski B Methods of roof cavability predic.source of energy and despite short-term fluctuations,tion. State of the Art of Ground Control in LongwallMining and Mining Subsidence, AIME, 1982: 13-29its long-term total demand in the world shows an [2 Afrouz A Hassani F, Scoble MJ. Geotechnical assess-pward trend. Reducing costs, achieving higher pro-of the bearing capacity of coal mine floors. Inter-duction rates per shift and increasing safety levels arenational Journal of Mining and Geological Engineering,1988(6):297-312and should therefore be studied carefully before final{aSn的的的巨Therefore mine mechanization is required. Mine [3]Bandopadhyay S. Partiathe most important problems in Iranian coal minesof primary strimechanization needs large amounts of investmentsBrown D, Singh R. An expert system for slope stabilityessment. Intemational Journal Surface Miningdecisions on the implementation of mechanization In1987:12-19our study, the most important parameters affecting [5] Denby B, Atkinson T. Expert system applications in themining industry. Mining Engineer, 1988: 505-514the viability of using mechanized mining of coal [6] Nguyen V. Some fuzzy set applications in mining ge-seams have been presented in terms of seam gradientmechanics. International Journal of Rock Mechanicsand thickness, geological disturbances, seam floorand Mining Science and Geomechanics, 1985, 22(6)conditions and roof formation. These parameters areimprecise. In our study by using fuzzy logic, mem- [7] Wenzhang G, Jingmin X Fuzzy evaluation of poweredsupport suitability to coal seams. International Journalbership functions and then fuzzy rule-bases are creof Mining and Geological Engineering, 1990(8): 163-ated and ultimately, the potential for mechanizationl77has been studied. Potential mechanization of the [8] Chen Y. Fuzy Set Theory. Wuhan: Central China Insti-Takht coal mines in Iran was investigated as a casetute of Technology Press, 1984: 154-159[9] Dubois D, Prade H. Fuzzy Sets and Systems: Theory andstudy. The results show a high potential for mechaniApplications. New York: Academic Press, 1980: 231zation in most of the takht coal seams[10] Kaufman A. Introduction to Theory of Fuzzy Subsets.New York: Academic Press, 1975: 45-132中国煤化工CNMHG