Martin Brook,Brue Helewhite,Rudrajit Mitra

a School of Environment,University of Auckland,Auckland 1010,New Zealand

b School of Mining Engineering,University of New South Wales,NSW 2052,Australia

c School of Mining Engineering,University of the Witwatersrand,Johannesburg 2050,South Africa

Keywords:Rock mass classification Roof strength Coal mine CMRR RMR Bowen Basin

ABSTRACT The rock mass rating(RMR)has been used across the geotechnical industry for half a century.In contrast,the coal mine roof rating(CMRR)was specifically introduced to underground coal mines two decades ago to link geological characterization with geotechnical risk mitigation.The premise of CMRR is that strength properties of mine roof rock are influenced by defects typical of coal measures stratigraphy.The CMRR has been used in longwall pillar design,roof support methods,and evaluation of extended cuts,but is rarely evaluated.Here,the RMR and CMRR are applied to a longwall coal mine.Roof rock mass classifications were undertaken at 67 locations across the mine.Both classifications showed marked spatial variability in terms of roof conditions.Normal and reverse faulting occur across the mine,and while no clear relationships exist between rock mass character and faulting,a central graben zone showed heterogeneous rock mass properties,and divergence between CMRR and RMR.Overall,the CMRR data fell within the broad envelope of results reported for extended cuts at Australian and U.S.coal mines.The corollary is that the CMRR is useful,and should not be used in isolation,but rather as a component of a strata control programme.

1.Introduction

The engineering design and stability of underground coal mines are longstanding issues,especially the gateroads that allow safe extraction of longwall panels [1].Failure mechanisms are dependent on the stress regime,as well as geological and geotechnical conditions[2].Roof and rib deformation typical of a longwall gateroad begins with the redistribution of in situ stresses around the new excavation [3].The impact of the stress changes on the roof and ribs is affected by the strength of the coal seam,and the underlying and overlying strata,as well as folding and faulting [4].The sedimentary layered rock that comprises a coal mine roof can fail in several ways,with roof sagging,and delamination of sedimentary layers often a precursor to failure [3].Moreover,after the extraction of a longwall panel,creating a void(goaf),rock fracture and stress redistribution occur,and so gateroads are subject to complex loading patterns[1].Hence,as within the civil tunnelling industry,rock mass characterisation schemes are often used in coal mining to aid engineering design and required roof support [5].

Of the range of different rock mass characterization approaches that have been introduced,the rock mass rating (RMR)in particular,has often been used in coal mining [6-11].The coal mine roof rating (CMRR)scheme was proposed specifically for underground coal mines,and has subsequently been applied over the last two decades,particularly in the U.S.,Australia and South Africa [3,12-17].The CMRR focuses on sedimentary layering,because the delamination is a key control on roof failure [12].Since its introduction,the CMRR has been used in a range of ground control issues.The CMRR has been used to constrain numerical modelling of rock mass deformation,as well as investigations for pillar design[18,19].The CMRR has also been used to assess roof bolt performance and roof failure[20,21].In several countries,CMRR has been specifically used to predict safe coal mining advance distances of extended-cuts [22].The CMRR has even been claimed to be an‘‘established coal industry standard”in Queensland coal mines,Australia[23].Despite this assertion,some workers have cautioned that the inappropriate application of the CMRR may have negative implications for strata control [24].Moreover,recent work highlighted potential issues with the reliability of CMRR in the Illinois Basin,U.S.,related to moisture sensitivity in weak roof conditions[17].Nevertheless,a recent review of the application of CMRR in Chinese coal mines found that CMRR is a useful method for preliminary investigations of roof stability [25].

While a broad body of literature exists comparing RMR with other established classification schemes such as rock mass quality index(Q),the CMRR has rarely been evaluated against other established rock mass classification schemes [10,26].Here,the CMRR data from 67 sites at the Carborough Downs underground coal mine,Queensland,Australia is presented in order to evaluate the applicability of the CMRR classification scheme.The RMR classification scheme is also applied in order to allow a direct comparison with the CMRR scheme.The aim in this paper is to investigate the applicability of CMRR in an underground longwall setting,and to determine whether rock mass conditions are closely related to geological structures.

2.Geological setting

2.1.Bowen Basin,Queensland,Australia

The 600 km long Bowen Basin contains an about 10 km thickness of clastic terrestrial and shallow marine sediments,that include petroleum and coal formations [27].The Bowen Basin is located in the northern zone of the Early Permian East Australian Rift System.Westward Late Permian-Mid Triassic thrusting of the New England Orogen caused foreland loading,leading to major subsidence and the accumulation of thick coal measures [27,28].During the Late Permian,major interfluvial peat accumulation occurred,prior to deposition of thick clastic sequences over the coal beds.This was followed in the Early Triassic,by uplift and erosion [29].

The focus here are the subhorizontal Rangal Coal Measures,which are in the Late Permian of central Queensland[28,30].These are the youngest formation in the Blackwater Group,have a spatial distribution of ＞20000 km2,and are overlain by the (Triassic)Rewan Group[29].Two regionally-extensive coal‘‘superseams”exist in the Rangal Coal Measures in the northern part of the Bowen Basin.They are the Leichhardt seam and the older Vermont seam[31].The study site is at Carborough Downs underground coal mine(Fig.1),150 km southwest of Mackay,Queensland,which extracts,via longwalls,coal from the Leichhardt Seam[32].The Leichhardt Seam at Carborough Downs ranges from 4.5 to 5.7 m thick,and consists of bright and interbedded dull coals.The mine produces predominantly coking coal for the international steel-making market[33].Roof and floor strength tends to be determined by grain size and therefore paleoenvironmental depositional setting.The roof strata often comprises carbonaceous mudstone in the first 0-5 m above the Leichhardt Seam,grading into stronger interbedded silts and sandstones.Hence,the roof can vary from weak to moderately strong (3-75 MPa with the mean of 30 MPa)rock,increasing to a maximum of 100 MPa in the presence of massive sandstone.The immediate floor strata underlying the Leichhardt Seam is typically mudstone,and considered to be weak,with a mean strength of approximately 20 MPa,and is as low as＜10 MPa in some areas[34].The mudstone is underlain by sandstone and siltstone layers,sometimes with carbonaceous partings,and is considered to be generally weaker than the roof.

2.2.Geological structures and in situ stress field

The complex structural history of the Bowen Basin since rifting in the early Permian,Mid-Late Permian subsidence,then Middle Triassic thrust faulting,is manifest in the structural geology [29].Indeed,a range of joint and cleat orientations and patterns are evident,reflecting both the present-day in situ stress field as well as previous deformation events [29].The Leichhardt Seam and overlying strata exhibits variable cleat and joint density,and with the existence of faults,this indicates that deformational events leave a strain record [35].The mine occurs within a shallow basinal structure east of the New Chum Fault,which has an estimated displacement of 100 m [36].While normal faults are formed around anticlinal structures,thrust faulting and associated bedding plane shears are also prevalent.Often,Permian normal faults have later been reactivated as reverse faults [37].While individual thrust fault segments may extend for c.3 km if N-S trending,shorter en-echelon style segments occur where the trend is NW or NE[29].In the northern sector of the Bowen Basin,there are three normal fault sets,trending NW,NE and E-W,with the northwesttrending normal faults particularly common [29].

The contemporary stress field characterised for the Bowen Basin displays a consistent north-northeast(007-027°)maximum horizontal stress orientation over 500 km [38].Measurements of in situ stresses at Carborough Downs Mine are consistent with the regional work,although locally the stress field can vary substantially around geological structures[38,39].Data has been reported for in situ stresses at 215 m depth in the western sector of Carborough Downs Mine[39].The mean values are:(1)maximum horizontal stress 8.9 MPa,bearing 30°,plunge 1°;(2)minimum horizontal stress 5.4 MPa,bearing 120°,plunge 8°;and (3)vertical stress 5.2 MPa.

The vertical stress magnitude(5.2 MPa)is slightly less than the nominal 5.4 MPa of overburden strata expected at 215 m depth with a mean density of 2500 kg/m3[39].

3.Methods

3.1.Site selection

Engineering mapping along underground gateroads identified a series of major faults within the mine,including both normal and reverse faulting,with throws of ＞3 m.These structures and associated discontinuities exert a key control on the roof and rib stability(Fig.2).Sites for rock mass description and characterisation were then selected on both normal and reverse fault-blocks,in order to provide scope for a range of discontinuity patterns.Face mapping data was augmented with discontinuity data obtained from drill core.The methods for determining the rock mass parameters,and the CMRR and RMR classification schemes are outlined below.A total of 67 sites were selected for CMRR and RMR classification.

3.2.Intact rock strength

The intact rock strength is an important parameter in both CMRR and RMR.Uniaxial compressive strength (UCS)was estimated via point load strength testing (PLT),following the established methods [40].Using the NX-sized core,the relationship between the Is50and UCS values were determined using the multiplier of 24 for 50 mm diameter specimens [40].Where PLT could not be used,UCS was estimated from sonic borehole logs,using the equation [41]:

UCS=882:07exp（-0:0381t） （1）

where t is the sonic transit time in microseconds per foot.This sonic-UCS relationship (R2=0.76)was established for Carborough Downs Mine,using 131 laboratory-determined UCS values,with strength ranging from 5.3 to 106 MPa [34].

3.3.Discontinuities,groundwater and weathering

From a geotechnical standpoint in coal measures rocks,discontinuous conditions are particularly important,including joint sets and spacing,along with roughness and any alteration or infilling of joints,in addition to groundwater effects[11].Both scanline surveys and drill core were used to determine discontinuities,with scanline joint data plotted in Rocscience Dips software.Pole concentrations were then contoured on the stereoplot [42].The average discontinuity spacing was determined for each joint set.Both butt and face cleats pervade coal seams at Carborough Downs Mine,and develop presumably from a multitude of mechanisms[35,43,44].However,they were rarely encountered here as the target stratigraphy for engineering was the roof rock overlying the Leichhardt Seam,which generally consisted of sand,silt and claystone.The rock quality designation (RQD)was estimated from the mean discontinuity frequency[1].Groundwater flow was estimated using descriptive classifications,while all sites were classed as ‘‘unweathered”[45].

3.4.Rock mass rating (RMR)

The RMR followed the established approach,whereby the five parameters of RQD,intact rock strength,condition and spacing of defects,and groundwater state,are then subject to a discontinuity adjustment [46].This then,provides a rating out of 100,whereby the rock can be classified using a five-point ordinal scale from‘‘very good rock”(RMR=81-100)to ‘‘very poor rock”(RMR ＜21).

3.5.Coal mine roof rating (CMRR)

The coal mine roof rating(CMRR)shares some similarities with the RMR in that it utilises both UCS and the discontinuity parameters,and that it is rated from 0 to 100,but it is calculated in two stages.This is outlined in detail,but for the sake of brevity,will only be summarised here[15].Initially,the mine roof is separated into different structural units,with unit ratings calculated for each unit,using Fig.3.A structural unit usually is one lithological unit,but several lithological units may be combined if the geotechnical properties are congruous.The unit rock mass rating is controlled by intact strength,defect intensity and shear strength,as well as several discontinuity adjustments,and moisture (Fig.3).This can all be determined from either underground exposures or from drill core.For the next stage,the weighted mean of all the individual unit ratings(which is affected by thickness)within the bolted roof is calculated as the CMRR index [15].The bolted horizon is the height into the mine roof that the bolts penetrate to (Fig.3).The effects of water,groundwater flow,and the nature of the contacts of different rock units are included in the CMRR that is finally calculated [15].

Fig.2.Roof conditions and geological structures at Carborough Downs Mine.

4.Results

4.1.RMR and CMRR values

In Table 1 are summarised the individual rock mass properties at each measurement site,including the RMR and CMRR total ratings values,derived from the individual parameters.Histograms of the frequency distribution of RMR and CMRR value are reported in Fig.4.The distribution of RMR data is a moderately peaked (approaching normal)distribution with the data centred around the mean of 56.4,and a standard deviation of 8.9 (Fig.4a).The values extend across three of the five RMR classes,with values focused within the ‘‘good rock”to ‘‘fair rock”classes.That the range of RMR values spanned three classes is indicative of the range of lithofacies encountered,from massive sandstone to coal measures.The CMRR values are more dispersed than the RMR values showing a mean of 57.1(standard deviation of±10.5),and a slight negative skewness.The vast majority of the data fall within the ‘‘moderate roof”class,although there are several evident outliers (Fig.4b)[15].

The distribution and range of RMR and CMRR values suggests that there is a moderate range of roof rock mass properties across the mine.This though,is unsurprising,because the measurements were all made across exposures and rock core of the Leichhardt Seam and associated sediments,rather than a single lithology,and results are not from a small restricted area within the mine.The variability in rock mass classification values observed is large enough to impact on mining continuity,hence,delineating and predicting these relatively moderate variations,and any spatial patterns in the distribution of rock mass classification values,is important.

4.2.RMR and CMRR variability and geological structures

The RMR and CMRR data were contoured using kriging,and superimposed on the mine layout and fault information,to facilitate comparisons between rock mass properties and structural geology.The contouring involved firstly defining the variogram parameters(linear variogram;no nugget;slope=1;no drift;anisotropy ratio=1,angle=0°;search radius=2700)to describe the edge of the area,prior to contouring.The RMR and CMRR contour maps are shown in Figs.5 and 6,respectively.In terms of the rock mass condition,as indicated by RMR and CMRR,there appears to be a distinction between the central zone of the mine around the south-central zones of longwall (LW)panels LW2,LW3,LW4,and LW5,and the northern and southern portions of the mine.Indeed,contours tend to show a much simpler pattern away from the marked heterogeneity,in terms of rock strength classifications shown in the central-south zone of the mine.

Fig.3.CMRR calculation based on the sum of individual unit ratings which extend to the limit of the bolted interval within the roof rock (after [12]).

The simple north-south contours define an elongate zone of high RMR between longwall panels LW2 and LW3 (Fig.5).This zone coincides with,and is constrained by,a series of northsouth graben structures,and extends northward to the northern end of LW5,bearing 10°,which closely parallels the direction of major horizontal stress (10°N).The highest RMR values tend to be associated with the downthrown (hangingwall)blocks of normal faults.The RMR pattern is also disrupted by faults to the west,where a zone of decreased RMR occurs to the west of LW1.Here,the zone of low RMR is bounded to the east by a reverse fault,so that the zone of weak RMR is positioned on the hangingwall block(Fig.5).

The pattern of the CMRR data (Fig.6)shows much more subdued strength ratings than the RMR data.Indeed,the LW3/LW4 area coincides with CMRR ratings of ＜48,defined as ‘‘moderate”to‘‘weak roof”[12].Hence,the weaker CMRR appears to be associated with the graben structures in the central area of the mine.A zone of elevated CMRR consistent with the high RMR occurs at the north end of LW5 (Fig.5).As with RMR,a broad zone of weak CMRR is located around the bord-and-pillar area at the south of the North Mains roadway (Fig.6).

A comparison of the CMRR and RMR data is reported in Fig.7,and this shows a residual contour plot (residual=CMRR-RMR)of the spatial pattern of difference between the RMR and CMRR values.As shown in Fig.7,both RMR and CMRR appear to closely reflect rock mass conditions,with a limited variability in the spatial pattern of CMRR compared with RMR.This close relationship is reflected in the linear regression relationship reported in Fig.7,which indicates a moderate to strong positive relationship between RMR and CMRR parameters (R2=0.46,p ＜0.001).

In summary,the rock mass classification data interpolated across the mine shows that three zones appear to be present that display contrasting rock mass conditions:

(1)Zone of roof rock mass strength complexity coinciding with graben structures between the LW2 and LW3 extraction panels(RMR=strong;CMRR=weak);this zone was deemed too problematic geotechnically to extract a longwall panel from;

(2)A weaker zone centred on the bord-and-pillar workings to the southwest of the mine,close to the drift (both RMR and CMRR);

(3)An area of high strength rock toward the north of LW5 panel,which coincides with the hangingwall of a reverse fault and the footwall of a normal fault.

5.Discussion

The key controls on roof deformation in underground coal mines are reported to be the in situ stress pattern and strength of the roof [1].This may result in the minor delamination or more vertically extensive block or bending failures in a low stress situation,while buckling failures may occur in a situation of higher stress [3].The results show that rock mass properties do vary across Carborough Downs Mine,as indicated by the varying geological structures that are evident,and the associated roof conditions.Nevertheless,a clear correspondence between mapped faults and rock mass classification values for CMRR and RMR is difficult to delineate with certainty.Encountering geological structures at Carborough Downs has important implications for mining during longwall panel extraction.Structure and defects in the Leichhardt Seam and overlying siliciclastic roof rocks havehad a number of problematic effects,including reducing the stability of roof and ribs in underground workings,and creating transfer pathways for the influx of water and gas into underground workings.Moreover,as shown at other underground coal mines,rock mass classification systems are an important stepping-stone from describing rock mass condition towards predicting rock mass

behaviour,and some of the implications of this study are discussed below [9].

Table 1 Summary observations of rock mass condition and derived rock mass classification values for CMRR and RMR.

Fig.4.Histograms of calculated rock mass classification values for RMR and CMRR.

5.1.Geological structures and rock mass condition

Delineating precise relationships between stress regime,faulting and roof stability in underground coal mines can be problematic [9].The contemporary maximum horizontal stress field(007-027°)for the Bowen Basin is consistent with reported situ stresses at Carborough Downs Mine[38,39].While seam dip direction follows this trend,it is perpendicular to both the strike of reverse and normal faulting,which is oblique to the maximum horizontal stress direction.Nevertheless,the faults are easily identifiable through bedding plane throws,and these faults follow the general structural trends (NNW-SSE)identified across the Bowen Basin[29].While reverse faults tend to follow a NW-SE strike,with minor exceptions,following the regional patterns,normal faulting around LW2 and LW3 appears to coincide with more complex roof rock strength [29].While these steeply-dipping structures are an important aspect of roof failure patterns observed at the mine,possible effects of faulting are eased somewhat by the general competency of the siltstone and sandstone roof overlying the Leichhardt Seam [30,34].

Notwithstanding the general roof stability and lack of an obvious relationship to faulting,roof failures can occur via the formation of guttering features,elongate roof cavities forming at the along the roof/rib interface [47].Although relationships between stress,faulting and the RMR and CMRR contour patterns were equivocal,guttering failures appear to be oriented at 90° to the maximum horizontal stress.At Carborough Downs,guttering failures take the form of canoe-shaped elongate cavities,oriented NW-SE along gateroads,often bounded by a joint plane toward the gateroad centre [48].Indeed,the stress orientation is usually oriented at 90° to the axis of gutters,thus the elongate,canoe-shaped gutters are analogous to a stress ellipse[49].Over time,the guttering can deteriorate until it results in a small roof failure.In the context of Australian coal mines,this observation is consistent with other datasets,whereby stresses tend to initiate guttering at the corner of roofs,by the deflection of roof beams as defined by defects [50].This also accords with other findings that guttering involves the gradual failure of brittle rock [48].

Fig.7.Residual contour plot of RMR value subtracted from CMRR value and scatterplot of CMRR and RMR,including linear regression best-fit and 95%confidence intervals about the regression line.There is no basis for considering either RMR or CMRR as independent(or dependent)of the other,so the designation of RMR as x and CMRR as y is made arbitrarily.

5.2.Rock mass classifications

Often,workers have advised the use of at least two classification systems when studying rock mass behaviour underground,and in this study,CMRR and RMR were used [46].Both the CMRR and RMR classification systems are sensitive to changes in rock mass properties.The two classification schemes generally show a close positive association,but this does vary and areas of geological complexity appear to cause a divergence in the relationship.While it is difficult to deduce how geological structures exact a control on the rock mass character causing the RMR and CMRR variability,intact strength likely has a strong control on the overall rating values,as it correlates strongly with both RMR (R=0.53,p ＜0.001)and CMRR (R=0.75,p ＜0.001).

The RMR values all fall within the bounds of values typically reported by other authors,with mean values of 56.4 and 57.1 for RMR and CMRR,respectively.The restricted range of values as indicated by the central tendency of the frequency distributions,indicates a modest range of rock mass condition across the mine.Regarding the range of RMR values,they are similar to those reported for the Late Eocene Waikato Coal Measures exposed in the Huntly East coal mine in New Zealand[9].There,the Kupakupa Seam is of similar extractive thickness to the Leichhardt Seam (＞5 m thick),and is of similar depth below the surface (150-300 m),with RMR values of 51-72.The Carborough Downs RMR value range (31-76)is also around the range of values (RMR=41-48)reported for an Indian underground coal mine [11].The values are also similar to the RMR range(37-57)reported from seven different coal mines across India [51].Hence,it would appear that RMR has been a useful parameter in rock mass classification at Carborough Downs,in that the RMR values are realistic and comparable with other studies.

Regarding the CMRR,the vast majority of published CMRR values originate from work in the U.S.by the original developers of the CMRR technique [12].This has led to cautioning on the use of CMRR in Australian coal mines without the use of a comparator classification scheme such as RMR [24].A related criticism is that CMRR was developed to characterise the roof conditions and roof support requirements of several underground coal mines mainly in the Appalachian area,that have moderately strong roof conditions[17].The corollary is that the use of CMRR for roof condition determination in coal mines with low strength roof units could be problematic.However,the use of RMR in parallel with CMRR,and the general competence of the siltstone and sandstone roof at Carborough Downs,indicates that CMRR is appropriate in an Australian coal mine setting.

The original developers of the CMRR suggest the following characterization of roof condition when using CMRR:(1)CMRR ＜45,weak roof;(2)CMRR=45-65,moderate roof;(3)CMRR ＞65,strong roof)[12].During their original U.S.research,75% of the data were in the weak or moderate categories;while in Australia,86%of collected CMRR data fell into the weak to moderate category[12,16].In comparison,79% of the Carborough Downs CMRR data fall within the weak to moderate roof categories,very similar to the previous work [12,16].Unfortunately,CMRR values often appear in published work without proper reference to the provenance of the original data,making comparison between different mines and coal measure sequences problematic [16].Hence,there are only a few selected studies that the Carborough Downs data can be directly compared with.Nevertheless,the range of CMRR values (28-81)calculated from Carborough Downs appears to accord with the range of CMRR reported from existing Australian and international coal mines.For example,the CMRR range is almost identical to the CMRR values (30-80)reported for Australian coal mines[24].CMRR values of 15-70 have been reported as typical for Australian coal mines,but the authors refrained from naming the mines,or citing specific data sources [16].It was also stated that the average CMRR for Australian longwall mines was 50,compared with CMRR=53 in the U.S.[16].Again,no specific sources are cited by the authors for the data,but the mean CMRR of c.57 from Carborough Downs is within these data ranges [16].CMRR values of 38-52 were reported for the roof above the German Creek Seam at Grasstree Mine in the central Bowen Basin,Australia[23].These are all within the weak roof to moderate roof as specified,and overlap closely with the CMRR values from the present study [12].In South African coal mines,CMRR values of 20-71 were reported from the Witbank,Free State and High Veld coal fields [13].Again,the Carborough Downs data closely overlap with these ranges of values.

Few studies exist that have attempted to compare CMRR and RMR from the same study site meaning that the Carborough Downs data are somewhat unique.One such example is the analysis of a highwall undertaken at Moura Mine in the southern Bowen Basin [52].However,the five RMR/CMRR data points that were reported indicated a negative relationship between RMR and CMRR,which is counter-intuitive,and contrasts with the findings at Carborough Downs.Indeed,the Carborough Downs data clearly indicate that a positive relationship should be identifiable if the sample size is sufficient.Hence,that the Moura Mine data show an inverse relationship is probably spurious,and a larger sample size(＞50)may produce a more representative relationship between RMR and CMRR[52].Nevertheless,at Carborough Downs between LW2 and LW3,high RMR coincided with lower CMRR,and this was associated with a series of faults,including grabens,and has remained unmined due to the geological complexity.

5.3.Application of CMRR to extended cuts

A final interesting comparison is the possible relationship between cover depth,geological structures,and failure mechanisms around mine gateways [53,54].Generally,cover depth part-controls the stress of rock around a gateway,and as stress increases concomitant with cover depth,large deformation zones occur in deep coal mines [54].A relationship between CMRR,roof stability and depth to the seam roof during extended cuts has been reported[16].Extended cuts in underground coal mines are situations where a continuous miner machine is advanced greater than 6 m‘‘inbye”the last row of permanent roof support[22].The relationship between CMRR,mining depth and extended cut stability from U.S.,Australia and the present Carborough Downs study is given in Fig.8 [16,20].Strictly speaking,the Carborough Downs data are not from extended cuts,but it is interesting to compare the Carborough Downs data,relative to the stability characterisations elsewhere [16,20].The discriminant trend-line is also included on the plot,and this is the line assumed to delineate the‘‘always stable”from the‘‘sometimes stable/never stable”roof examples [20].The assumption is that the higher the CMRR,the more likely development excavation underground will be successful,and that the roof behaviour remains static,depending,in part,on the cover depth (and in situ stress direction and magnitude).Whether this is a valid assessment is open to conjecture,however,88%of the Carborough Downs CMRR values all plot above the discriminant line,suggesting general roof stability [20].Hence,this implies that the CMRR is a useful indicator of roof stability.Moreover,the discriminant trend-line appears to be an appropriate guide to the likely divergence from static behaviour to roof buckling,delineating a requirement for increased roof support requirements during extended cuts [16].

Fig.8.Relationship between CMRR,depth and the stability of extended cuts,using data from U.S.and Australia [16,20].The Carborough Downs data is also superimposed on the plot.

6.Conclusion

The coal mine roof rating (CMRR)and rock mass rating (RMR)rock mass characterization methods have been used to illustrate roof conditions for an underground coal mine situated in the Bowen Basin,Queensland,Australia.Both techniques have shown consistency over large parts of the mine,with divergence in overall ratings in areas of geological complexity.Nevertheless,although identification of major structures gives an indication of rock mass character complexity,the exact connection between structural geology and rock mass classification values remains equivocal.In summary,this study shows that:(1)the CMRR system is a useful approach for characterizing roof rock mass condition;(2)the intact strength exerts a key influence on both CMRR and RMR,and thus is an essential parameter for roof classification and roof support design;and(3)the applicability of the CMRR system for prediction of roof stability in extended cuts closely accords with international datasets.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

Staff at Vale Australia,in particular Lachlan Cunningham and Priscilla Page,are thanked for facilitating underground access to the Carborough Downs Mine.The research was kindly supported by Moultrie Group and Golder Associates.