Australian Coal Geology Volume 3 Part 2 1981 - CGG by GSAustralia

CONTENTS Volume 3

Part 2

Australian Coal Geology

Colin Ward: Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

Ray Williams and Robin Turner: An interpretation of the mining conditions at Stockton Borehole Colliery by geological analysis . . 63

Journal of the Coal Geology Group of the Geological Society of Australia

J.Shepherd, J.W.Creasey and N.I.Fisher: Statistical analysis of spacings between geological discontinuities in coal mines, with applications to short-range forecasting of mining conditions ... .. . . .. .. ......... . ........ . .. . ................ . . 71

Ian Stone: Regional and local vanatlons in v1tnmte reflectance patterns and their application to strata control investigations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

P .P. Wootton: Computer application m exploration and planning of Hunter Valley No. I Mine . ... . ........ . .. . 91

Bulent Agrali: Application of geology in longwall mining a case history of longwall 202 in the Bulli seam at South Bulli Colliery . . . .... . .......... .. .. . . 97

Graeme Boyd: Coal measure geology and the planning and operation of open strip coal mines ............ . . ... 107

J.R.Enever and R.M.McKean: Techniques for the determination of geomechanical parameters during surface exploration ......... 117

Volume 3

G.R.Mcllveen: Engineering geology and the environment - do coal geologists have environmental responsibilities?. 135

1981 Part 2

Coal Geology Group of the Geological Society of Australia

Chairman: Vice - Chairman: Hon. Secretary: Hon. Treasurer: Hon. Editor:

Colin Ward Ian Blayden Michelle Smyth Richard Britten Anton Crouch

Editorial panel:

Anton Crouch, Colin Ward

Address for editorial matters:

Address for other matters:

Anton Crouch 300 Glebe Point Road , Glebe, N .S.W. 2037

Laurel Tuckwell Geological Society of Australia, Room 1001 , Challis House, 10 Martin Place, Sydney, N .S.W. 2000

ISSN 0725 - 4938 Price to non-members, $ 10.00 per issue

Australian Coal Geology

Journal of the Coal Geology Group of the Geological Society of Australia

Volume 3 1981 Part 2

Papers from the symposium "Progress in coal geology", Sydney, 26 - 28 August 1980

Volume 3

Part 2

Colin Ward: Int ro d uction . ....... . . ..... .. . . ... . . . .. . . .... . . .. .. . .. . ......... . .. ..... .. . .. . .. .. .. .. . . . . .. . 61

Ray Williams and Robin Turner: An interpretation of the mining conditions at Stockton Borehole Colliery by geological analysis .. 63

J.Shepherd, J.W.Creasey and N.I.Fisher: Statistical analysis of spacings between geological discontinuities in coal mines, with applicatio ns to short-range forecasting of mining conditions . .. .. ........ .................... .. .. .............. 7 1

Ian Stone: Regional and local vanat1ons in v1trm1te reflectance patterns and their application to st ra ta control investigations .................... ........ . .............. . ........ . . . .. . . . ... . .... . .... 8 1

P.P.Wootton: Computer application in exploration and planning of Hunter Valley No . I Mine ................. 9 1

Bulent Agrali: Application of geology in longwall mmmg a case history of longwall 202 in the Bulli seam at South Bulli Colliery ... . ..... . .. . ....... . ... 97

Graeme Boyd: Coal measure geology and the plan ning and operation of ope n strip coal mines .............. . .. 107

J.R.Enever and R.M.McKean: Tec hniques fo r th e determ ina ti o n of geo mechanical parameters d uring su rface ex pl o ra ti o n ....... .. 117

G.R.Mcllveen: Engineering geology a nd th e enviro nme nt - d o coal geologists have environ mental respo nsi bili ties?. I 35

Issued , April 1983

Introduction

Colin Ward Coal Geology Group Chairman, 1979-81. The eight papers in this issue of "Australian Coal Geology" repre ent the second instalment of those presented at the 1980 symposium on "Progress in coal geology". Heavy demands on people's time during the subsequent period of unprecedented activity in the coal industry have delayed their publication to some extent, and indeed another Coal Geology Group symposium has now taken place since that time. Following on from the first six papers published from the "Progress" ymposium, which dealt with recent advances in geophysics for the coal industry (Volume 3, Part I), the papers in this issue are devoted to developments in mining geology, geomechanics and computer-based geological data processing. The Australian coal industry is currently a world leader in these fields, and Australian coal geologists will undoubtedly become more involved with them as exploration of coalfield areas passes into mine design and operation. These skills and technology will also be needed as both old and new projects strive to maximize their efficiency and maintain their competitive edge in both the domestic and overseas market place. Once again , thanks are due to Anton Crouch and Ian Blayden for editing of the manuscript~ . Th e authors of the various paper are to be thanked for their contributions, and , in particular, for th eir co-operation with the editors' requirements in getting the papers ready for publicati o n.

Mr Richard Austen giving the keynote address at the symposium

Part of the technical exhibition at the symposium. (The three figures are not part of the exhibition but are geologists looking at some of the material on display).

An interpretation of the mining conditions at Stockton Borehole Colliery by geological analysis Ray Williams 1 and Robin Turner 2• I. 2.

Collinsville Coal Company, P.O. Box 60, COLLINSVILLE, QLD 4084. B.H.P. Co. Ltd, P.O. Box 196B, NEWCASTLE, N.S. W. 2300

ABSTRACT The mining conditions in the Young Wallsend seam of the Newcastle Coal Measures range from fair to very difficult. Strata stability is generally poor, fundamentally because of the presence of an incompetent bed of tuff overlying the seam. This tuff has been extensively deformed by tectonic events, and consequently provides little resistance to the shear stresses induced by mining. On a more regional scale, between panels or between areas of a panel, variations in strata stability are due to changes in the magnitude of the virgin stress field, particularly variations in the degree of stored strain energy. Geological reasons for this stress heterogeneity are described . A good correlation has been established between the dip of the north-west striking cleat and the degree of strata stability. The cleat set' in itself does not affect stability, but is used as an indication of stored strain energy that may be present. Kinematic and kinetic analyses have been shown to be of fundamental importance in understanding the variations in strata stability. The most recent tectonic event in the area is an indicator of the direction and relative magnitude of the principal stress axes for the stress field that existed before mining took place.

INTRODUCTION Stockton Borehole Colliery is owned by the Broken Hill Proprietary Co . Ltd. It is situated at the northern end of Lake Macquarie (Figure 1), with workings in the Young Wallsend seam. This seam has been worked under generally difficult roof conditions. Failure of the roof strata at the face has had a significant, adverse effect on production while time-dependent post-mining failure has led to a high degree of behind-the-face maintenance. To assist in overcoming this problem, geological and rock mechanics investigations were carried They were aimed at understanding the out. controls on strata deformation and were seen as a basis for the design of subsequent quantitative studies and changes to elements of the mining system. At present, an alternative roof support system at the mining face is being implemented, which is proving more effective than previous methods in controlling the difficult roof conditions encountered. The analysis for this paper is restricted to first workings in a bord and pillar operation.

tlond• kGreto

,'---,,,_=,_,:;~ District

REA COV ERED Y I

i1londWot1o9o Coal Distr

. Coo l

---4--

I C!

VERE O

y 2

t,TAS

tie Cool

CENTRAL SOUTHERN BELT

!ne t

COALFI ELD

160

120

Kll OMETI/ES

1. COALFIELDS OF THE SYDNEY BASIN

NEWC AS TLE

GEOLOGY Sedimentology The Young Wallsend seam lies within and near the base of the Newcastle Coal Measures. The lower 2.3m to 2.6m of the seam is mined, leaving from 0.3m to 0.6m of coal in the roof. The roof coal is of inferior quality and contains numerous thin clay

Kt l O ME Tl?ES

2. COLLIERY HOLDING IN THE NEWCASTLE COALFIELD

Figure I

64 bands and partings (Figure 2). The Young Wallsend seam is overlain by the obbys Tuff Member, which maintains a uniform thickness of 1.6m over the colliery holding. In vertical section, the tuff exhibits a wide variation in composition and strength, ranging from soft montmorillonite horizons to hard siderite/ chalcedony horizons. obby's Tuff Member 1s overlain by The approximately 1.5m of dark grey fossiliferous mudstone with coaly shale and coal bands. The combined 3.0m of strata overlying the Young Wallsend seam are laterally very uniform. Above these, however, the strata vary and tend to coarsen in the direction of a major north-east trending palaeochannel system, which is situated near the north-western colliery boundary (Figure 3). The floor strata of the Young Wallsend seam, although lithologically more variable than the roof strata, are mainly of sandstone throughout the worked-out area. They do, however, become dominantly shaly near the boundary with the Macquarie Colliery to the south. This change in lithology reflects the existence of a north-west trending palaeochannel system between the Young Wallsend seam and underlying Yard seam in the area of the western most workings.

Structure Stockton Borehole Colliery is located on the axis of the Macquarie Syncline, which plunges to the south at IO (Figure 3). Over most of the workedout areas, the seam dips uniformly to the southsouth-west at 2° . In the western area of the workings, however, the seam changes attitude across the synclinal axis, where it dips towards the east at I O • The depth of cover varies from 240m in the north to 300m in the south. The dominant fracture pattern comprises a series of north-west striking normal faults , dykes, cleats and joints. Two zones of normal faulting have been proven, the maximum throw being 5.2m (Figure 4) . There are two zones of north-West striking dykes (Figure 3). One of these occurs just beyond the south-western extremity of the colliery workings. The other dyke zone has been intersected in a number of places by the colliery workings, and, at its widest , was found to be 40m across with individual dykes ranging up to 2m thick. Since this dyke zone is the only one accessible, subsequent reference to a north-west striking dyke zone relates to this feature . The term #cleat# is used in this paper to denote geological fractures, other than faults, occurring in coal, while the term #j oint# is used to denote geological fractures other than faults , occurring in the non-coal strata. The north-west striking cleats are sub-parallel to the north-west striking normal faults and dykes, and have an average strike of 319° (I.S.G.). These cleats are well developed and are often penetrative for the full seam thickness. orth-west striking joints are also developed, but to a relatively minor extent, and strike at 337° (I.S.G .). A further structural trend is defined by an east-

GR AINSIZ[

LlhlOLOGY

-~:-:-

VAR IATIONS •ltf.Mft.

L\/llr(

1 -+--+---+---"--<,__...,....__,...._.......--r--, ----+-,-

I I - ' -+----t--+-,1-+--,--0-- - t - - - t - - t - --t- - - •

irirriTf

~11ic,r.1e

I~~~

_ 11~ 1

LEGEND

-.:I ~

Coot Sholy Cool

Cooly St'IOi.

IIlllIJD ~ ~

s ,1111one

CJ Figure 2

striking dyke and an east-striking cleat set. The dyke bisects the colliery workings, and is approximately 19m thick (Figure 3). The eaststriking cleats are non-penetrative, rarely extending laterally for more than 100mm. They have an average strike of 103° (I.S.G.). A third structural trend is defined by two eastdipping thrust faults and associated shear fractures. These faults have a northerly strike and a throw of up to 1.8m. Only the westernmost fault is accessible in the mine workings, and subsequent reference to a thrust fault therefore relates to this particular structure. An extensive system of low-angle conjugate shear fractures is developed in the Nobbys Tuff These fractures generally have a Member. northerly strike and are genetically related to the thrust faults.

MINING CONDITIONS In conjunction with geological analyses, mmmginduced strain features were mapped at a scale of I :500. All signs of failure and many indications of stress and stress direction were noted . The type

Figure 3

and sequence of mining were recorded, including the roof support method used . The floor and rib conditions were generally very good, with almost all adverse deformation occurring in the roof strata. Thus, reference to strata deformation herein, implies involvement of the roof strata only. A study of the empirical data indicated two basic modes of roof failure. The first involves failure at or near the working face , and usually takes place in the first eight hours after mining. The second type of failure occurs weeks or even months after mining has taken place, and is referred to as timedependent failure . For failure at or near the face, variations in mining conditions occurred on a "local" scale and a "regional" scale. Variations on a "local" scale occur within and between mine roadways, while "regional" variations occur between panels, or between areas of a single panel. The roof mapping program almost invariably showed that the headings experienced a significantly higher degree of roof deterioration than the cut-throughs. The headings commonly experienced poor roof conditions, whereas the cutthroughs experienced fair to good conditions irrespective of the mining direction. The failure that did take place in the headings and cutthroughs, however, was generally found to be biased towards one side of the opening and often There was a strong occurred as guttering. tendency for roof failures to occur down the side of the roadway in which the north-west striking cleat was first exposed during mining (Figure 4) .

lhC' norl.h wnt stroking dtot

-,,

E.01.tuty odYor,u d,rc-ct ,011 - quttt• n,g

Witst,rly ochonc:, d 1rttt,on - 91-11· , r·nq on t.OUthtrn

on no<ltwrn 1,,c:,

1,oc 01 09C" ·"9

...,

Chr,c1oon - qu t'"' "'l

of opcn ,ncJ

A '

D•OI

root

opcr11ng

::.OUt N rly odvo· ,

Nortlwrty o<lvon.., d ·rtttoon-q.,tt,r,ng

fr1:11lur ,

NoMn • O\''' ·I

M)Ulh .., <! s'• •t

d 11T1,.:,o"

odvor,.cc

b,01 o1 root

IO,lfOrO

• lhrr 1,d r o1

Figure 4

lolwr r 10 .. or llll

r0011..-o,

66 Variations in mining conditions on what is termed a *regional scale* are illustrated in Figure 5. The mining conditions have been delineated into two basic types, poor and fair. Two main types of poor conditions were mapped . The first was where failure of the roof took place at the face during mining, and where a high degree of rooffloor convergence occurred soon after mining. In many instances, however, the roof convergence continued with time, and after a period of weeks or even months, roof falls still occasionally occurred . Falls of this second type were quite different to those occurring at or near the face . They were often up to 8m in height, whereas the falls occurring around the face were rarely greater than Im high.

INTERPRETATION Failure at or near the face Failure at or near the face, on both a "local" and "regional" scale, was found to result from the interaction of high lateral virgin stresses in the strata with geological discontinuities, modified perhaps by the mining process. The fact that headings almost always experienced a greater degree of roof deterioration than cutthroughs was found to be controlled by the sequence of mining. In the continuous mining method, headings are the first roadways to be dri ven into the virgin coal. The openings deform under the imposed stress and create a zone of stress relief about them through which cutthroughs are driven. This type of failure can only be promoted by high lateral stresses. If vertical loads predominated, the later-formed roadways would experience additional loading in the abutments of the previously driven roadways. Lateral compression failure is widespread and is easily identified . Lateral shortening of the roof strata is commonly observed and is often measured in tens of centimetres. The fact that roof failure generally occurs down the side of the roadway in which the north-west striking cleat is first exposed , indicates that the cleat controls the way in which the lateral stress is relieved . Differential movement along cleats would result in some stress relaxation ahead of the face on the side opposite the initial intersection of the north-west striking cleat (left hand side in Figure 6) . On the side in which the cleat is first exposed, the stress concentration occurring at the corner of the opening is not relieved by movement along the cleat but is, instead, relieved by fracturing of the coal roof itself. Areas of progressively deteriorating roof conditions (*regional scale* failure) were found to be due to high lateral stress concentrations, apparently promoted by: (a) residual tectonic stresses associated with the thrust faulting; (b) the north-west striking dyke zone; (c) an area of converging palaeochannel systems in the roof and floor strata. A zone of poor mining conditions was always associated with the thrust fault (Figure 5). Within

:· --- - ·"'\_~ ·--,,... ,

· ·• J

'·-----,

_____ _.,,,,/

i -~ t,[l··

':., .,' / jjf

•-:::;,:;~, (i"-""•

LEGEND

i .

S TOCKTON B

=--=--~• • c-."'

.; Mc..c,;, o ;;;~;:ala;a--~;&.:_:~ ~- - - ~

" - - , _ , _ , . 10 1. , ...,."' c- -

COLLIE

fw-....- '"

~ --

Cft!Nrool-

,......,_"""° ___ ,,,_ --a _,....,.,.INl'llflq, ... to"'4fl....,...OI

...

...,__- ,'"'"",,....._.,

SCALE I Z!IOCX>

Figure 5

The coal is relat ively distressed because relaxat ion is assisted by different ial movement along the cleats.

~ .?' North - west striking deat

The coal is more highly stressed because the cleat cannot assist in the release of strain energy.

I MINE

RO._, (Pton \Atw)

Figure 6

Guttering

67 is horizontal and In a virgin stress field in which has a magnitude of (say) three times the vertical load, very high lateral stresses are induced in the immediate roof strata and a low-profile linear arch is produced (Figure 9) . Because the confining pressures are low around the periphery of the opening, high shear stresses are generated within the arch . They are greatest at the corners of the openings where the confining pressures are at a minimum. The shear strength of the roof coal is generally less than the shear stresses, and failure of The arch the linear arch results (Figure 9). progressively migrates upward as the strata within it fracture, placing the strata beneath the arch under vertical tension (Figure I0).

this zone, the north-west striking cleats exhibited a steeper dip than cleats with a similar orientation located in areas of relatively good mining conditions. In the zone of poor mining conditions the dip of the cleat generally ranged from 88° SW When relatively good mining E. to 87° conditions were experienced , the dip ranged from 82° SW to 87° SW (Figure 7). It is postulated that the more steeply dipping cleats were rotated in response to a force couple generated during the formation of the thrust faults. The rotation, albeit by only a few degrees, is indicative of an increase in the degree of stored strain energy which is released upon mining and leads to the poor roof conditions. A zone of poor mining conditions was experienced on the north-eastern side of the north-west striking dyke zone. After penetrating through the dyke zone, conditions suddenly improved. The nature of roof failure on the north-eastern side of the dyke zone indicated very high lateral stresses with lateral shortening of up to 0.3m. It was quite apparent that the dyke zone had acted as a barrier, promoting stress concentration. The poor roof conditions over the westernmost workings are interpreted as resulting from a total stress concentration occurring in the vicinity of major palaeochannels in the roof and floor strata. This is thought to be brought about by the lenticular geometry of the palaeochannel systems and the low modulus of the coal seams (Figure 8).

0' 1

POOR

MINING

CONDITIONS

GOOD

MINING

CONDITIONS

Time-dependent failure Olp of

Time-dependent failure is explained by means of a model that utilises the linear arch concept of Woodruff (1966). The linear arch describes the nature of the mining-induced stress field which is, to a large extent, influenced by the nature of the virgin stress field.

The norl h wesT

itrilu n9

Cl@OI

Figure 7

High Modulus Material (sandstone and conglomerate)

High Modulus

NOTE :

The area

of stress

Material

concentration is depicted by the hatched area

Figure 8

The arch continues to migrate upwards because the clayey na ture a nd low a ngle geological shear planes in the Nobbys Tuff Member often causes its shear strength to be less than th e shear stresses generated in the a rch . As th e arch mrgra tes upward , the shear stresses in it progressive ly decrease, and a stable arch configuration is reached when th e shea r stresses in the a rch ha ve a magnitude less than the shear stre ngth of th e stra ta forming the arch (Figure 11). This co nfigu ra t ion is often reached weeks or months after mining. Roof stability then depends upon the a bili ty of th e roof supports to resist the dead weight of strata beneath the arch line.

Mud stone Tu ff

Undisturbed cond1t1on

Mud stone

Shearing adjacent an arch abutment due to high lateral stress acting on weak strata

Figure 9

Mudslone

Roof bolts are not long enough and installed early enough to half the continued upward m1gra11on of the arch into strata above the Nobbys Tuff.

Cool

Upward m ig ration of the arch is accompanied by a reduction in the level of lat eral stress in the arch and the changing of the immediate roof strata stresses to vert ica lly tensile

I I

•>

Tull

Coal

MU<:lstone

Cool

A stable arch configuration 1s finally reached when the height of the arch causes a reduction in the level of stress to below the yie ld point of the strata in whi ch II occurs. Stresses in the Nobbys Tuff have been rotated to near vertical resulting in a reduced tendency for shear failure .

Figure I I 2.

Shear failure occurs along pre -exi sting planes in the tuft. The roof 1s v1s1bly sagged and the arch con tinues to migrate upward.

Figure JO

CONCLUSION The generally poor roof conditions at Stockton Borehole Colliery are the result of high lateral The stresses acting upon incompetent strata . regional variations in mining conditions indicate a heterogeneous virgin stress field , the heterogeneity The being caused by geological structures. mining-induced stress field is similarly variable, and is significantly affected by the sequence of mining. Variations in heading direction have had no discernable effect on roof stability. The immediate three metres of roof strata are laterally uniform in lithology and it is found that changes in mechanical properties do not contribute significantly to the failure pattern . obbys Tuff may have Weathering of the contributed to time-dependent failure, but there is no direct evidence in support of this. Time-dependent failure results from progressive Th is upward fracturing of the linear arch . mechanism is thought to be generally applicable to most mining situations experiencing timedependent failure.

ACKNOWLEDG EMENTS The Broken Hill Proprietary Co. Ltd . is acknowledged for permission to publish this paper. The views expressed are those of the authors and do not necessarily reflect those of the company.

REFERENCES Woodruff S.D., 1966,Methods of working coal and metal mines, Yol I - roof control and support problems, Pergamm on Press, Oxford.

r--

Statistical analysis of spacings between geological discontinuities 1n coal mines, with applications to short-range forecasting of m1n1ng conditions J.Shepherd 1*, J.W.Creasey 1 and N.I.Fisber 2 CS/RO Division of Mineral Physics, P.O. Box 136, NORTH RYDE, N.S. W. 2113 2. CS/RO Division of Mathematics and Statistics, P.O. Box 218, LINDFIELD, N.S. W. 2070 • Present Address: Australian Coal Industry Research Laboratories Ltd, P.O. Box 83, NORTH R YDE, N.S. W. 2113 1.

ABSTRACT Black coal seams are affected by geological discontinuities that form recognisa ble sets, and which are characterised by the properties of orientation, shape, length, vertical continuity and spacing. Whilst the properties of orientation and length have been q uite extensively investigated, the varia bility of spacings of discontinuities, especially in coal, is largely unknown. Measurements of discontinuity spacings in the coal seams of several coalfields have been made using the scanline technique, particularly around fault zones that have resulted in deleterious mining conditions. The spacings along a scanline can be analysed to dis tinguish between segments of equally spaced discontinuities, segments exhibiting Poisson behaviour and segments containing (possibly periodic) clusters. A substantial increase in the intensity of occurrence of discontinuities has been found in the vicinity of fault zones . Thus identification of increased discontinuity intensity as mining proceeds can be used to forecast approaching fault zones and give warning of any associated unstable roof conditions or areas of gas outburst-prone coal.

INTROD UCTION One of the principal pro blems of underground coal extraction is the need for geological information concerning the fault and joint (cleat) pattern in reserves earmarked for mining. The disruption to mining caused by drivage into unpredicted tectonic structures is well known (see for example Clarke, I 966; Shepherd and Fisher, I 978). However, surprisingly few techniques are available to mining geologists and engineers for the assessment of these structures prior to or during mining. The analysis of coal joint spacings p1'esented in this paper is based on recent studies aimed at finding reasons for the recurring bad mining conditions encountered in the Lithgow seam of the Western Coalfield of New South Wales. In this coalfield, roof conditions can deteriorate very suddenly during advancement into virgin coal. In a distance of less than !Orn from good conditions, further drivage may only be possible by the erection of a very closely spaced pattern of roofbolts, W-straps, mesh, legs and bars. The association of such bad roof conditions with faults and joints is described elsewhere (Shepherd and others, 1980). Empirically, it has been found that large numbers of joints of a certain set occur in swarms of widths varying from 50m to 200m, and that where the joint freq uency is greater than

4 or 5 per metre, a potentially unstable roof ex ists. Clearly, there is a need for techniq ues that can assist in forecasting the location of such swarms before mining reaches them . Collieries provide extensive and almost continuous exposures of a coal seam, together with the traces of tectonic structures traversing the working Thus ideal conditions exist for the section. measurement of the spacing of structures. Very little data on discontinuity spacings has been recorded (Cobbold and Ferguson, 1979), and therefore analytical techniques for the treatment of such data are virtually unknown. In this paper, preliminary analyses are presented of coal joint spacings along scanlines in two collieries in the Western Coalfield. For this analysis, the joint positions are regarded as a point process (i.e. a sequence of points haphazardly positioned along a line). As a means of identifying clusters, the real joint spacings are compared with artifically generated spacings data of known properties. The most important aspect of the analysis is that spacings are studied as an ordered sequence along the scanline, rather than as a collection of values without regard to their position (Cox a nd Lewis, 1966).

JOINT SPACINGS AND DOMAINS Systematic, parallel and sub-parallel joints form sets (Price , 1966) and the perpendicular separations of adjacent joints of the same set are termed the spacings (Wheeler and Dixon, 1980). Schematic joint spacing in a typical colliery roof is shown in Figure I. Joint frequency is the inverse of spacings, expressed as the number of joints per unit distance. Joint intensity is the sum of the joint surface areas (m 2) per unit rock volume (m 3). Comprehensive descriptions for the quantification of discontinuities have been given by the International Society for Rock Mechanics ( 1978) and Wheeler and Stubbs ( 1979). The present authors have studied two functions computed from the spacings of coal joints, termed the "joi nt spacings intensity" and the "joi nt spacings moving coefficient of va riation (MCV)" (see Appendix). A few authors have considered fracture spacings in rocks, including Price ( 1966) who showed that variations are found with bed thickness. Thamm (1 939) mapped joints in the sediments of the Pretoria Series in South Africa and discovered that an increase in frequency occurred close to major faults (Figure 2) . Recently , and independently from the present authors' studies, Murawski (I 979) has shown that certain joint systems in coal seams are associated with major faults. Macrae and Lawson (1954) measured cleat spacings i n Yorkshire coal seams and demonstrated some variability of spacings according to the nature of the coal ply. Ting (1977) and Takranov ( 1979) also considered cleat spacings in relation to rank and coal type. Engineering geologists have used joint spacings semi-quantitatively and the various classes have been reviewed by Watkins (1971). A more detailed study of discontinuity spacings from scanline data has been made by Priest and Hudson (1976) and Hudson and Priest (1979). They showed that the distribution of spacings values derived from scanline traverses, when plotted as frequencies of spacings classes, generally approaches a negative exponential form of curve. These studies were carried out in relation to the need for characterising rock mass behaviour for engineering purposes. However, these authors did not consider orientation (i.e. set) or length and did not explicitly consider the distribution of spacings as being derived from a point process along a scanline. Joints distributed along a scanline may conform to the point process categories shown in Figure 3. In coal seams, the natural joint distributions are probably a combination of these processes and therefore analytical techniques are needed to distinguish them. In particular, statistical techniques are required to examine joint spacings along a scanline for: (i) the occurrence of closely spaced joints forming clusters, and the magnitude of these clusters; (ii) the variation in the pattern of joint spacings along a scanline. It is particularly important to establish the type of background (or natural) process for joints if

/som6fric ◄

Yl~W Of

roadway

APPARENT SPACING

Diagrammatic representation of the spacings of joints along a colliery roadway.

Figure

"-' a: ,"-' X:

ISO

"-' 0.. <I)

!: 100

SITE Of fAIA.T

.... 0

a: ,o "-'

ID :,: :::,

800

600

,oo

200

,oo

600

0 I STANCES '""""'

Figure 2

Frequency of joints near a fault in the Pretoria Series of South Africa (after Thamm, 1939).

SYSTEMATIC

EQUAL SPACING ( no clusters)

JOINT SET

~ UNEQUAL SPACING ~

NON-POISSON (clustering inexplicoble by o random process e.g. periodic clustering)

Figure 3

RANDOM or POISSON (irregular clustering)

Possible point processes that joint locations form along a scanline.

deviations from that process are to be detected . Studies so far indicate that some faults are associated with abnormal joint clustering, so it is essential to have a means of recognising changes in the process due to such clustering.

73 Blocks of coal that are characterised by homogeneous joint and / or fault patterns can be regarded as structural domains (Figure 4) . The nature of a structural domain has been shown in a number of coalfields to be related to roof conditions (Moelle, I 978 ; Shepherd and others, I 980), and fault swarms, such as those described by Shepherd and Burns (1978) , are a domain especially deleterious to mining. Recent studies aimed at improving the identification of structural domains have been carried out by Huntington and Fisher ( 1980).

Figure 4

MEASUREMENT OF JOINT SPACINGS ALONG A SCANLINE Spacings of joints of the same set are measured by using the scanline technique (Piteau, 1970). A measuring tape is set up near the roof or along the rib at easily identifiable points. Where a coal roof exists and there is no obstruction from support, a scanline traverse in plan can be carried out continuously along a roadway. Otherwise, rib traverses are used , but these usually lack continuity due to the presence of cut-throughs. A schematic plan of a scanline is shown in Figure 5. For scanlines along mine roofs, the centre-pins at intersections are used as traverse endpoints. In Figure 5, for example, only joint traces with lengths greater than 3.0m are measured in the scanline. Coded data sheets are used for recording joint spacings. The data is stored for subsequent processing in a "spacing data base file" of a minicomputer system.

STATISTICAL GENERATION SPACINGS DATA

Initially, the properties of the joint spacings intensity and the moving coefficient of variation (MCV) were studied by generating artificial data with known characteristics and plotting intensity and MCV . Three types of spacings distributions were generated: (i) spacings from an exponential distribution, corresponding to a Poisson point process (sometimes referred to as a random point process); (ii) almost equally spaced data; (iii) clustered data. These data sets were then merged to produce the following combinations: (i) random and clustered data; (ii) almost equally spaced and clustered data.

STATISTICAL ANALYSIS OF SPACINGS Joint frequency Variations in joint frequency are determined by plotting a histogram of the numbers of joints in unit intervals (classes) along a traverse . If the joint spacings are apparent, then the histogram is a plot of apparent joint frequency. Class sizes are variable, but usually in the vicinity of 0.5m to I.Orn.

Diagrammatic representation of structural domains in a coal seam based on the orientation and spacing ofj oints. Domains I and Ill are typical of good mining conditions, and Domain II is typical of bad mining conditions). PLAN

J l Figure 5

L I

SCANLINE

Schematic plan of a scanline survey in a colliery roadway roof (Joint traces are classified by length. In the example given, joints labelled • I" were measured and those labelled •r were rejected).

Joint spacing intensity The use of classes (unit intervals) in analysing joint spacings is unsatisfactory as the classes interfere with the spacings of adjacent joints at class boundaries. Therefore, a "smoothed" version of spacings treatment is required, that is not affected by changes in the location of class intervals . The "k-nearest-neighbour" technique has been used in the analysis of joint spacings to produce values of the intensity function (see Appendix for details). The intensity function is then plotted at equally spaced points along a traverse (see Appendix for some sample calculations). It highlights areas of closely spaced joints (clusters) along the traverse. Clusters form peaks that vary in magnitude depending on the number and spacing of joints within the cluster. Filtering for cluster size can be carried out by increasing k, the nearest-neighbour number. Clusters consisting of few closely spaced joints will be reduced or merge into the background intensity with increasing values of k, while higher intensity clusters will only be reduced marginally.

Moving coefficient of variation A test for exponentiality is used to try to identify the type or types of processes that joint spacings form along a scanline traverse. The type of processes that possibly exist are outlined above . The technique involves calculating the mean (x) and standard deviation (s) of joint spacings within a group of joints; group size is usually greater than 20, so that satisfactory determinations 9f x and s are made. •

74 CV is recalculated and plotted at the midpoint of the new 20 - spacing interval. This is repeated until the last spacing is encountered . By increasing the number of spacings used to calculate the CV at each stage, a smoother plot is obtained which may give a better determination of the nature of the underlying process and in fact determine if there is a change in behaviour along a scanline. However, the drawback of this can be the smoothing out of small (periodic) clusters.

In a Poisson process the distributions of spacings is exponential , and the mean and standard deviation are the same. Therefore, if the coefficient of variation (CV) of a sample of spacings is approximately I, there is an indication that the spacings have an exponential distribution and hence that the locations of the joints constitute a Poisson (random) process. If CV - 0, then the joints tend to be equally spaced; if CV> I, then the joints are probably from a point process with a different clustering nature (e.g. periodic) from that expected with a Poisson process. To perform the moving CV plot, the CV is computed for the first (say) 20 spacings and its value plotted at the midpoint of the 20 - spacing interval (as shown in the Appendix) . The first spacing is deleted and the 21 st added in, then the

ANALYSIS OF SIMULATED SPACINGS DATA Exponential and clustered joint spacings The computer plots of this data are shown in Figure 6. The raw data plot (Figure 6A) shows the

Sl"ULATED CLUSTERS ON RANDO" DATA All11 UTH Of

TRAY[AS[ •

90.00 DEGREES

RZIHL,IH Of SET

0.00 OlGR[[S

NO. OF SPACltiCiS • 209

!'ilAqm.o~ 0.0

30 . 0

20.0

10.0

ITJD~l,,n,Jo,rldl~lli~&~m m,onOllq O O:rml n

ffi OID fl J 40 . 0

50.0

CLASS I NT EAVAL •

70.0

60.0

BO.a

90.0

100.0

11 0.0

1?0.0

130.0

Jl,IO . n

MET RE S

15 1'l .0

I. 00 NE TAE 151

METRES

NEAREST NE JGH80UR Of S

I NTEAVAL• D. SO NETAE

JO I NTS

2.0 2. St

u >

J.5

1.0

z >

o.s

0. 0+-~---,-~~---,-~-,----,-

0. 0

10 '. o

Jo'. o

20'.o

110'. o

-- 11ETRES ,---,---r-.---r--r---.--r--~---.-~-,----,-,---,---r-.---r---r-.---r--r---. 1 1 1 1 1

sc'.o

so '. o

CV VALUES FOR GROUPS Of 25

Figure 6 A. B. C. D.

10'. o

eo'. o

go'. o '100 '. o

110'.o

120 '. o

130 '. o

140 '. o

1so '. o

JOINTS

Analytical plots of clusters superimposed on exponentially distributed joint spacings. Schematic plan of joint spacings. Histogram of joint frequency. Plot of joint spacings intensity function. Plot of moving coefficient of variaion (MCV).

75 sequence of exponential spacings together with The clusters are three superimposed clusters. spaced 15m apart and each contains 20 joints. The joint frequency is given in Figure 68 as a histogram. Joint frequencies vary from a bout I to 5 per metre for the randomly distributed joints away from the clusters and up to 10.5 per metre in the joint clusters . The plot of the spacings intensity function in Figu re 6C provides a better definition of the clusters than Figure 6B. Some clusters occur in the randomly distributed joints at various positions, for example at 27m and 55m from the starting point of the plot. These clusters are inherent in a random point process and only attain an intensity of approximately half that of the superimposed clusters.

In the MCV plot of Figure 6D, the spacings for groups of 25 joints away from the superimposed clusters suggests an approach to an exponential distri bution where the values of the function are Around the superimposed clusters, the unity. spacings di stribution appears not to be exponential, with values of the function greater than unity.

Merged equally spaced and clustered joints (exaggerated example) Figure 7 shows computer plots of the same three simulated joint clusters superimposed on almost equally spaced joints. The raw data is shown in Figure 7A. The joint frequency is given in Figure 78. The smoothed joint frequency is about I per metre for

Sl"ULATED CLUSTERS ON ALNOST EQUALLT SPACED DATA AZIH UTH Of

TRAV ERSE • 90.00 DEGREES

AZJ HUTH OF SET • 0.00 DEGREES

NO. OF SPACINGS • 209

JO 9 8

1 6

•

3 2 I 0 +u-'"'t'.u.;~4"-"f'w..l.f.lLl"'t'-''"-'t....t'-""+-"'-'-'tlll.lt'-"u+'LU.j-'W'f-U-"f-l-"-YU.U.4".LI.fJ.W.f.1.LL'fUJ..l.fLil.1f.Ll..l.l.fl.Jw.+llJ..1f-LU'4'--Lllf.llUifil'-4' illjIso . o JYO . o 130 . 0 120 . 0 11 0 . 0 100. 0 9 0.o ea.a 10 .0 so .a so.a so .a 30 .o 20.0 10. 0 a.a

CLASS I NTERVAL •

NETRE S

I. 00 "ETRE 151

1- 2.0 2 . 51 ,..

"' ~ 1.5 z ~

1.0

0. 5

NE TRE S ~-~~~-~~~-~...,...~-~------~-_-__-r---~---:_-~---_--,_-_-:_,-;_-_-~-----~------;:~-~...,...~-~~~~~ ~ a. o-+--:.--,..-_-:.,-;_ I I I I I 1 1 1-_ o. o

10 0

20'.o

Jo'.o

oo'.o

INTERVAL •

so'.o

11 0 0

so'o

0.50 "ETRE

NEAREST NEIGHBOUR OF 5

120 '.o

130 '. o

100'.o

Iso 0

JOINTS

NE IRE S

Figure 7

CV VALUES FOR GROUPS OF 25

J OINTS

Analy tical plots of clusters superimposed on almost equally spaced joints. (For explanation of plots A - D, refer to Figure 6).

76 almost equally spaced joints, rising to between 9 and 11 per metre in the joint clusters. The joint spacings intensity plot in Figure 7C shows that the superimposed clusters are very clearly defined in contrast to the much lower background intensity. The MCV plot is shown in Figure 7D; the MCV value approaches zero in the background regions. In the vicinity of the superimposed clusters, the values of the function increase abruptly.

required heavy support. Joint spacings measurements were made along the scanline shown in Figure 8.

••

SWARM

'""-/

Effects of clusters on background patterns This analysis has only considered superimposed clusters of 20 joints with a width of 4m. The effects of different sizes of clusters in background distributions are currently under investigation. In the two cases considered here, the spacings intensity function has clearly highlighted relatively small clusters. The effects of the superimposed clusters on the two types of background processes are to increase the value of the standard deviation relative to the mean for the groups of spacings along the traverse. Thus, changes in the nature of the point process along a traverse have been identified by using the MCV.

MARGIN OF FAI/II 8 ,KJ/NT ;

• SCANLIN£

OI/TLJNE ~

ROOF f:J RA>FRSE

OF WORKINGS

CONDITIONS

KJOm

•-~ t f

Figure 8

Outline plan of Wallerawang No. 3 Mine workings, showing a fault and joint swarm, as well as the location of a scanline traverse. (Arrows indicate direction of drivage).

ANALYSIS OF JOINTS IN PROXIMITY TO FAULT ZONES. Two actual joint spacings data sets have been analysed from the Wallerawang and Ivanhoe Collieries in the Western Coalfield. Both data sets originate from geological structures that were mining obstacles at the time of the development of the first workings. Actual joint spacings have also been measured in other coalfields, particularly at West Cliff Colliery, in the Southern Coalfield of N.S. W., in connection with strike-slip faults prone to instantaneous outbursts. Shepherd and Creasey ( 1979) found that marked increases in coal joint frequencies occurred commonly at distances of 45m from a strike-slip fault and well before outburst-prone conditions occurred. A point along a scanline traverse at which a sustained increase in joint frequency is identified is termed the prediction threshold and this can be used to obtain some forewarning of a fault immediately ahead of an advancing face . It is emphasised that the present authors' joint spacings studies have concentrated on regions in the proximity of faults and fault swarms and that, in the course of this work, it has become apparent that wide domains devoid of mine roof instability and away from faults may provide the key to an understanding of the background joint process(es).

Wallerawang An outline of a portion of the first workings of the now disused Wallerawang No.3 mine is shown in Figure 8. Access to reserves was jeopardised by the occurrence of very difficult roof conditions in the main entries. Structural geological studies described elsewhere (Shepherd and others, 1980) defined the western margin of a minor fault and joint swarm . The Lithgow seam is continuous through this swarm, with displacements of less than 0.5m, but the highly jointed coal roof

The analytical plots of the Wallerawang 200-metre long scanline are shown in Figure 9. The occurrence of minor normal faults , as well as bad roof conditions, is shown in Figure 9A. The joint distribution is variable, with a few wide spacings and some isolated low-intensity clusters (Figure 9C). For this portion of the traverse, the MCV plot shows values close to unity, possibly suggesting a Poisson process for the joint locations . Such joint spacings may represent the background conditions found away from fault swarms. The margin of the fault swarm at approximately I !Orn is associated with a high-intensity cluster and high variable spacings (Figure 9C and 9D). This cluster resembles the superimposed clusters merged on exponential spacings in the simulated data set (Figures 6A to 6D). The sustained increase in joint frequency between 140m and 200m is associated with roof failure and a number of closely spaced, moderate to highintensity clusters. Values in the range of 0.5 - 1.0 for the MCV in this segment confirm that clustering is present. The observation that a high-intensity cluster occurs at the margin of the fault swarm (Figures 9A and 9C), at a distance of 45m from the onset of bad roof conditions, suggests that this fault and joint swarm could have been forecast if drivage had been from the west towards the east. Since faults are virtually absent at locations remote from swarms in the Western Coalfield, it is suggested that high-intensity clusters may be indicative of the margin of a fault swarm. A further point is that the sustained increase in joint frequency and recurring minor faults provide a warning of impending difficult roof conditions.

77 WAI.LEMINING NCI., NINE SCANLIN£ AZIMUTH OF TARV(RS[ •

122 . 8] DEGREES

ORTA ARE J OINT PLANES

AZINUTH OF SET • 22.52 DECREES

NO. OF SPACI NGS • 11143

·11~, ,U.., ,L~·~••LJw...~ ijLl J..J ,,1' o.o

10. 0

lo.o

"'o .o

so . o

&o.o

10.0

CLASS INTEflYAt. • I.DO 11£TAE 151

JNTUIJAL• 0 . 5D "![HIE

Figure 9

20 .0

YAl.UES fM Cfll0Ul'5 Of 30

NEAMST

ao.o

90.0

N( 1GHIOUft

100.0

Of 5

110.0

120.0

1:,0.0

1110.0

1so.o

160.0

110.0

1ao .o

190.0

JOINTS

Analy tical plots of the Wa/lerawang scanline survey. (For explanation of plots A - D, refer to Figure 6).

MARGIN OF ir-FAULT ZONE

PRINCIPAL

Ivanhoe fault zone During development of the main entries at Ivanhoe Colliery, a normal fault zone with a seam displacement of 5.2m was met. The workings were reduced to three headings to penetrate this zone (Figure 10). A cross section of the fault zone and slight monoclinal flexure is shown in Figure 11 . This cross section is drawn along the same heading as that in which the scanline traverse was made (Figure 10). The eastern segment of the scanline was rotated to compute the apparent spaci ngs along a straight line across the fault zone. The first 30m section of the scanline has a low joint frequency and some low-intensity clusters, and indicates a single process with MCV values of less than unity (Figure 12). The appearance of more intense clusters from the 40m mark has shown up in the MCV plot as a change in the underlying process. From 80m to 100m along the traverse, sustained higher joint frequencies with high-intensity clusters occur across the principal fault zone. On the eastern margin of the fault zone a similar pattern of low-intensity clusters occurs, which changes eastwards to a lowfrequency, cluster-free process that is probably the background condition .

\ \

--- .. - I

:I '-

•--

.-. •

-'SCANLINE TRAVERSE

........ 0

Figure 10

IOOm

Outline plan of the Ivanhoe Fault zone and location of the scanline survey.

78 EAST

WCST

Cross section of the Ivanhoe Fault zone (CT is cut-through).

occurrence of the thrusts on the western side and the normal faults on the eastern side. The early warning signs that the workings are approaching the fault zone may be the presence of low to medium-intensity joint clusters at a distance of 22m from the main fault in the footwall block and 43m in the hanging wall block. The presence of these joint clusters is considered to be comparable with the marginal cluster at a distance of 45m from the western margin of the Wallerawang fault swarm.

The principal fault plane occurs at the western margin of the high-frequency joint swarm (at 82m along the scanline) and the large numbers of joints are coincident with several minor faults to the east in the hanging wall block . The appearance of low to medium-intensity clusters on each ide of the fault swarm probably indicates the presence of the fault zone and as such may be of predictive value, e pecially if recognised in conjunction with the

Scanline surveys carried out to measure Jomt spacings in colliery roofs or ribs should take into account joint sets and individual joint length. The locations of joint positions along a scanline can be regarded as constituting a point process and algorithms have been devised to assist in the identification ~f joint clusters. Joint spacings

OIRCCT>ON OF DRfVAGE OF ltlAJN ROADWAY

ZONE OF MINOR THRUSTS

------'

ZONE OF NOHMAL FAULTS

SCAM AND COLL/CRY FUJOR SCALE

Figure I I

l!.m

CONCLUSIONS

IVANHOE FAULT ZONE SCANLIN£ RZ J" UT'i OF TRAVERSE • 80.50 OC:Cifl£ES

• Allt'UIH OF S£1 • 3511.00 OEC~HS h0. OF SPAC I.\CS • 21.o S

OPTA R~E JO i "il PLA"iES

ill/I Biil II llll lml 11n1m~ Hffllll■■ll■Hll•m II Ill III Ill"r' I """""""""""""""""""' J

""·~·."""""""""

"'"

"""""""

[

MAIN tA~~·ZONE

:o e

•

;

~ ~ ~

! : I 0

NET AES

0.0

10.0

Z0.0

30 .0

li0.0

60.0

70.0

80. 0

90.0

100.0

110.C

120. 0

130.0

1'10.0

150 .0

160.0

CLASS INTEftYAL • 1.00 METftE 151

N[Aft[ST N[IGHBOUft Of S

JOINTS

METRES

Figure 12

CY YALU[S fOft GftOU~S Of 50

JOINTS

Analy tical plots of the lnvanhoe Fault zone scanline survey. For explanation of plots A - D, refer to Figure 6.

79 intensity plots of nearest-neighbour values for small groups of spacings readily define the positions and· magnitudes of joint clusters. Joint spacings along a scanline can be analysed for segments with equal spacings, segments showing Poisson behaviour and segmenl.s containing an unusual amount of clustering. This analysis has the potential for extracting much more information about joint spacings behaviour along a scanline than the previously utilised method of grouping spacings to compute frequencies , without regard to their behaviour as an ordered sequence. Changes in the joint distribution in coal seams have been found around certain types of fault zones, such as fault and joint swarms in the Western Coalfield and strike-slip faults in one part of the Southern Coalfield . The only advantage of a histogram over the intensity function appears to be that it gives an immediate reading of the number of joints per metre (if Im is the basic histogram interval). However, this quantity can also be obtained from the intensity function as the inverse value of the intensity at any given point. Apart from this, the intensity function does not suffer from the difficulty of the arbitrary location of class boundaries (as does the histogram). Fault and joint swarms in the Western Coalfield consist of high-intensity, closely spaced joint clusters, together with faults distributed as individuals and in clusters. It appea rs that three classes of joint cluster can be tentatively identified in the scanline traverses analysed so fa'r: highintensity clusters in a swarm; low to mediumintensity clusters associated with the margins of a swarm; and low-intensity clusters that are a part of the inter-swarm joint distribution . Future work on joint spacings will aim to determine the patterns of joint spacings in regions unaffected by faults or fault swarms and , in addition, to confirm or modify the tentative recognition of different cluster types.

ACKNOWLEDGEMENTS Blue Circle Southern Cement Ltd and Coalex Pty Ltd are thanked for permission to publish data from their collieries. We are grateful for the financial support provided by Austen and Butta Ltd , Blue Circle Southern Cement Ltd , Coalex Pty Ltd , the Electricity Commission of New South Wales, and the National Energy Research , Development and Demonstration programme administered by the Commonwealth Department of ational Development and Energy. Dr Jon Huntington wrote the original version of the computer program that plots the joint spacings, and we have benefited from generous discussions with him during the course of this work. Dr Peter Diggle of the Department of Statistics at the University of Newcastle-on-Tyne (U .K.) provided much useful input to the statistical aspects of this study and his help is gratefully acknowledged . Tony White is thanked for drafting the figures.

REFERENCES Clarke, A . M., 1966. Pay-off v. risk assessments in planning and mining geology, Min . Engr, 126, 29 - 46. Cobbold , P.R. and Ferguson, C.C. , 1979. Description and origin of spatial periodicity in tectonic structures: report on a Tectonic Studies Group conference held at ottingham University, 8 November 1978, J. Struct. Geo/., I (/). 93 - 97. Cox, D. R. and Lewis, P.A .W., 1966. The Statistical Analysis of Series of Events, Methuen, London. Hudson, J .A. and Priest, S. D., 1979. Discontinuities and rock mass geometry, lnt. J. Rock Mech. Min . Sci. & Geomech. Abstr. , 16, 339 • 362. Huntington, J .F. and Fisher, .I. , 1980. The identification of structural domains from geological orientation data, in Proc. 42nd lnt. Statist. Inst., Manila, Philippines, 4 - 14 December 1979 (in press). International Society for Rock Mechanics (I.S . R. M.) Commission on standardisation of laboratory and field tests, 1978. Suggested methods for the quantitative description of discontinuities in rock masses, lnt. J. Rock Mech. Min. Sci. & Geomech. Abstr., /5 (6). 319 - 368. Macrae, J .C. and Lawson, W., 1954. The incidence of cleat fractures in some Yorkshire coal sea m , Trans. Leeds Geol. Ass., 6, 227 - 242. Moelle, K.H.R ., 1978. Directional mining in roc ks of the Sydney Basi n, .S. W., in 3rd Aust. Tunn elling Conf, Sydney, 12 - 15 September. The Institution of Enginee rs, Australia, 7 - 10. Murawski, H., 1979. Coal seams as tectonic bodies, Z. dt. geol. Ges., 130, I • 14 (in German). Piteau, D. R. , 1970. Geological factors significant to the stability of slopes cut in rocks, in Van Rensburg, P.W.J. (Ed.), Proc. Symp. Planning Open Pit Min es, 33 - 53, A.A. Balkema, Ca pe Town. Price, N.J., 1966. Fault and Joint Development in Brittle and Semi-brittle Rock, Perga mon Press, Oxford . Priest, S.D. and Hud son, J .A. , 1976. Disco nt inu ity spacings in rock, lnt. J. Rock Mech. Min. Sci & Geomech. Abstr., /3, /35 - 148. Shepherd , J . and Burns, K.L. , 1978. Fault swarm · in the Greta coal seam, ew South Wales, Proc. Australas. Inst. Min. Metal/. , No. 267, 27 - 36. Shepherd , J . and Creasey, J.W., 1979. Forewarn ing of fa ults and outbursts of coa l and gas at West Cliff Colliery, Australia, Colliery Guard. Coal lnt., 227 (4), 13 • 22. Shepherd , J . and Fisher, .I. , 1978. Faults and their effects on coal mine ro of fa ilure and mining rate: a case stud y in a New South Wales colliery, Min. Engng, 30, 1325 - 1334. Shepherd , J ., Huntington, J . F. and Creasey, J .W., 1980. Surface and underground geologica l prediction of bad roof conditions in collieries of the Western Coalfield, New South Wales, Australia , Trans. Inst. Min . Metal/. (Section B. Applied Earth • Science), 90, I • I3.

80 Takranov, R.A ., 1979. Jointing in the coals of the Irkutsk Basin, Geolgiya i geofizika, 20 {I). 75 - 82. Thamm, ., 1939. Certain methods of mapping faults in the Pretoria and Dolomite Series on West Witwatersrand Areas, Ltd, Trans. Geol. Soc. S. Afr., 42, I - 13.

Ting, F.T.C., 1977. Origin and spacing of cleats in coal beds, in Energy Technology Conf. and Exhibition, Houston, Texas, September 18 - 22. The American Society of Mechanical Engineers, Paper No. 77 - Pet-24, 3 pp. Watkins, M.D., 1971. Terminology for describing the spacing of discontinuities of rock masses, Q. JI Engng Geol.. 3. 193 - 195.

Wheeler, R.L. and Dixon, J.M., 1980. Inten ity of ystematic joints: methods and application, Geology, 8. 230 - 233.

Wheeler, R. L. and Stubb , J.L. , 1979. Style elements of systematic joints: statistical analysis of size, spacing and other characteristics, in Podwy ocki, M. H. and Earle, J.L. (Ed ,), Proc. 2nd lnt. Conf. Basement Tectonics, Newark, Delaware, /976, 49/ - 499.

APPE DIX Joint spacings intensity using nearest-neighbour technique The technique requires measuring the distance to the kth nearest joint from a given point on a scanline. The inverse of twice this distance is the intensity function (IF) of pacings at that point, given by:

IF(x) = 2dk(x) where x = a rbitrary point on scanline, and dk(x) = di tance to kth neare t joint from x. If the fir tor la t joint of a scanline is encountered before the kth neare t joint is found from x, the kth neare t joint become the kth joint from that end of the sca nline . The endpoint rule then applie in calculating the inten ity function: t the start of a scanline, IF(x)

= (j ~-) k J1

At the end of a scanline, 1 IF(x) = (j_. ) n Jn-k + I

were j 1, .. ., j 0 are the joints along a scanline. For example, suppose the first 10 joint locations along a scanline are: 0.1, 0.25, 0.6, 0.85, 1.2, 1.8, 2.5, 2.7, 3.0, 3.6 For x = 0.5 and the nearest-neighbour number k = 5. the first joint (0.1) is the 4th neare t neighbour and the 5th joint must be 1.2. Thu , the endpoint rule applies and the intensity function I/ ( 1.2-0. 1) = 0.91. For x = 1.0, the 5th neare t joint is 1.8 and then :

IF(x) =

I = 0.63 2( 1.8- 1.0)

Coefficient of variation (CV) For any numbers, a" 8i ... a 0 CV=-=a

I n where a = -La.1 n i= l and

j J_ ~ (a.-a) n-1 i = I ,

The moving coefficient of variation (MCV) for a canline is calculated by calculating the CV of pacing within groups of joints, and plotting the value at the midpoint of the group. By incrementing one joint at a time, the group move along the scanline. Hence, the MCV for a particular group ize i determined along the canline. By u ing the above example, and a group size of 5 joints, the first MCV is for 0.1 , 0.25 , 0.6, 0.85, 1.2; wit h pacing of0. 15, 0.35, 0.25, 0.35, thus: a = 0.275 = 0.055 MCV = 0.201 Midpoint = 0.65 The econd MCV calculation is for the group 0.25, 0.6, 0.85, 1.2, 1.8; with spacing of 0.35, 0.25, 0.35, 0.6, thu : a = 0.388 = 0.086 MCV = 0.222 Midpoint = 0.775 In practice, at least 20 joints are required to obtain a rea onable determination of the behaviour of the pacings, and ometimes 30-40 joints should be used ( ee text) .

Regional and local variations 1n vitrinite reflectance patterns and their application to strata control investigations Ian Stone Clutha Development Pty Ltd, P. O.Box 12, NARELLAN, N.S. W. 2567

ABSTRACT The optical indicating surface of vitrinite obtained from oriented coal samples may be related to the lateral component of tectonic strain. Subsequent strain events may be recorded in the vitrinite and several strain direction peaks may commonly appear in the lateral strain "ellipse" because of incomplete overprinting. These strain patterns can be recognised for both regional and localised features. Localised variations in the regional strain field around normal faults, for examP.le, show a consistent pattern which is associated with stress build-up prior to faulting and restricted zones ot stress relief and re-orientation after faulting. Similarly, marked changes in the vitrinite reflectance character are recognised around reverse and strike-slip faults. Investigation of regional straip patterns in the Bulli sel!,m in Burragorang Valley coal mines, N .S. W. was undertaken from biaxial vitrinites with reflectances as low as 1.00%. Apart from perturbations of strain due to local features, two areas have been delineated which have different regional strain orientations. These two areas also have different roof stability characteristics when mined, and to some extent have different structural features. Vitrinite appears to be capable of recording strain events as recent as the uplift and unloading phase in the tectonic history of the .area. Analysis of the optical properties of vitrinites may be useful in adding to the geological information retrievable from borecores, and may be particularly useful as an aid to structural interpretation in relation to an established data base.

INTROD UCTI ON Recent studies of the non-uniaxial behaviour of vitrinite reflectance show that it may be an indicator of strain (Stone and Cook, 1979 a,b). In particular, the strain pattern around some normal faults, exposed during underground mining, has been measured in oriented samples of the Permian medium volatile bituminous coals of the Bulli seam. To date these strain patterns have only been measured around relatively small scale geological structures (normal faults with throws Jess than 60m and strike-slip faults with displacement less than Im), in the mildly deformed geological environment of the Southern Coalfield of the Sydney Basin (Figure I). Geological descriptions of this area are given by Wilson and others ( 1958), Mayne and others ( 1974). Such mildly deformed strata do not possess readily interpretable or apparent indicators of geological strain, especially over small distances (tens to hundreds of metres) . A more regional interpretation of the structure, however, may be attempted from fracture patterns (Shepherd and Huntington, 1981), in conjunction with the relatively gentle flexuring. The existence of directional horizontal stresses, with magnitudes greater than expected from lithostatic conditions in the Southern and South-Western Coalfields has been demonstrated from both stress measurements (Worotnicki and Denham, 1976; Blackwood and others, 1976; Enever and McKay, 1980) and the

behaviour of roof strata within the coal mines (Connelly, 1970; Williams and Wilson, 1976). The effective "locking in" of palaeostress events as residual strain within the rock mass (Friedman, 1972; Varnes and Lee, 1972; Voight, 1974) and subsequent interaction with, or modification of, any currently applied tectonic stress is unknown. Existing physical methods used to measure in-situ stress in the rock mass (Worotnicki and Walton, 1976) are only able to determine the net stress at that site. However, the observed release of strain energy from the roof strata in restricted areas of the Burragorang Valley mines, for example (Figure I), does challenge the idea that "locked in" palaeostress provides an insignificant part of the net stress field . A present day regional tectonic stress field would be more likely to have a ubiquitous distribution over any given area . From data gathered (Stone and Cook , 1979 a,b) it appears that vitrinite is able to record at least some of these palaeostress events. Other techniques used to detect strain in rock fabrics such as quartz microfractures (Roberts, 1965), analysis of quartz optic axes (Riekels and Baker, 1977), dynamic analysis of calcite twin Jamellae and quartz deformation lamellae (Burger and Hamill, 1976; Gale, 1980) and x-ray analysis of quartz lattice distortions (Friedman, 1972), have not been shown to be successful in such relatively undeformed strata. Although the non-coal strata in this study have no apparent penetrati_ve strain

reflectances they measured (Figure 2). The existence of bireflectance in the bedding plane section (a-b plane, Figure 2) of the biaxial indicating surface was thought to have resulted from an applied lateral stress. Subsequent studies have been directed to measuring the azimuth of the true maximum reflectance (Ro max), which coincides with the b-c plane (Figure 2), because this is thought to represent the maximum strain orientation which apparently developed normal to the a pp lied lateral stress. Cook and others ( 1972) suggest that the asymmetric growth of the polyaromatic micelles in the vitrnite, in response to a lateral stress, might account for the nonuniaxial behaviour of the optical indicating figure.

COALFIELD WOLLONGONG N.S. W.

IYDNt:Y .r"

L,, ■AUi ~

UNIAXIAL INDICATING SURFACE OBLATE SPHERO ID

;

BIAXIAL INDICATING SURFACE OBLATE ELLIPSOID

S Y DNEY

""-,

Figure 1

Locality plan of the study areas within the southern Sydney Basin. A - Darkes Forest & West Cliff mine leases; B - Burragorang Valley mine leases

a•b)c

features, the release of elastic strain upon mining, causing difficult roof conditions in some areas, suggests that the strain history is at least significant with respect to underground mining. This paper describes the results of palaeostrain measurements from vitrinite of the Bulli Coal, Burragorang Valley, .S. W. (Figure I). Samples were collected in four different mines in an attempt to gain a regional picture of the palaeostrain variation. There are three main aspects to the design of this study. Firstly, the itrinite reflectance of the Bulli Coal in this area is approximately 1.00% (McLean and Wright, 1976), which is lower than in coals previously studied (Stone and Cook, 1979 a,b) . Secondly, to compare the localised strain pattern of geological structures, such as faulting , with the variation of a regional strain pattern. Thirdly, the mines sampled contain two distinct areas which have very different mining conditions - the poor mining conditions being related to roof failure induced by a dominant lateral stress (Nicholls, 1978). All azimuths refer to the ISG north.

THEORY AND METHOD The shape of the optical indicating surface which describes vitrinite reflectance anistropy is used to interpret strain. A more detailed explanation of the theory and method of determining palaeostrain is given by Stone and Cook (1979 a,b), and is given briefly below. Normally the optical indicating surface is considered as an oblate spheroid (Hevia and Virgos, 1977), but work on anthracites (Cook and others, 1972) suggests that an oblate triaxial ellipsoid (biaxial indicating surface) might more readily explain the

Cl) a

Figure 2

b) a ) c

I I

ED· a

Idealised v11rm11e reflectance indicating surfaces for optically uniaxial vitrinite and optically biaxial vitrinite, after Stone and Cook (1979a)

The most practical way to determine the azimuth of maximum reflectance (used interchangeably with "strain direction") was to measure the mean maximum reflectance (Ro max) from six oriented polished sections which had been cut normal to bedding (assumed ho[izontal) from one oriente_d coal sample. The Ro max values and their respective orientations are used to construct the shape of the a-b plane of the indicating surface (Figure 2). This plane is more correctly termed the "calculated bedding plane section of the indicating surface", but is referred to herein as the "strain figure". The maximum reflectance direction is determined by using data from the six sections, in combinations of three, to calculate apparent Ro max directions. Their dispersion about the mean orientation is tested to establish whether the results are randomly distributed or if the vector means have an orientation which is statistically non-random. Typically, the strain figures with the smoother elliptical shapes have non-random Ro max orientations. Work to date indicates that the strain figure with a randomly oriented Ro max commonly has two distinct reflectance or strain peaks. Each peak of the bedding plane section of the indicating surface is considered to be a potentially separate strain event, and can be resolved into its main strain components for

83 interpretation. The accurate estimation of the strain components is necessarily restricted by the small number of sections which define each peak but if enough samples are taken from the different sites measured, the method allows the regional strain orientations to be estimated on a stati tical basis.

L O CAL ST R UCTU R ES - EXAM P LES FROM PR EV IO US WOR K Before variation of the regional strain field can be assessed, the extent and nature of the modifications of strain a bout localised structures, such as faults , should be known. The stresses associated with the formation of faults, such as normal faults, are well established and provide a gauge for testing the vitrinite strain results . Two examples of the interpretation of the strain figures and the nature of the strain variation around fault are presented below. Both examples are from the Bulli Coal, Southern Coalfield , .S. W. (We t Cliff and Darkes Forest Collieries - Figure I). The idealised strain pattern about a normal fault (Figure 3) has been constructed from the study of a number of such faults , two of which were detailed by Stone and Cook ( 1979 a). Up to three different strain events have been interpreted , each being defined by its angular relationship to the fault. The strains appear to be recorded on both sides of the fault. For the example shown in Figure 3, non-random orientations of maximum reflectance (strain ellipses A,C and E) are marked by the small bar and are significant at the 90% level. Ellipses D and B have random reflectance orientations. The three different strain events may be classified as Types I to 3 as follows: Type I (ellipse A, Figure 3) is the most distal from the fault and the Ro max does not necessarily have a fixed angular relationship with the fault. It is thought to represent the more regional strain; Type 2 (ellipse C, Figure 3) has an Ro max oriented greater than 60° to the fault direction and is indicative of the strain developed prior to faulting, that is, the maximum lateral stress has a similar trend to the fault - as is true for normal fault formation ; Type 3 (ellipse E, Figure 3) occurs immediately adjacent to the fault and is oriented within 35° of the fault direction. It is thought to represent strain developed by post-failure re-orientation of the major lateral stress on the fault plane. Ellipses D and B are transitional zones and. show the peaks of strain events from both the immediately adjacent strain types . This suggests that, where strain events are superimposed , both may be recorded in the vitrinite until the earlier event is completely overprinted. The second example, a more complex situation, is of two normal fault planes which have been The displaced by a dextral strike-slip fault. movement of these structures is less than 0.5m (Stone and Cook, 1979 b). The strike-slip fault is one of a family of structures, of similar orientation which have been locations of gas outbursts (Marshall and others, 1980).

F Figure 3

Generalised strain pallern that might develop around normal faulting . Individual strain figures (A-£) may have non-random Ro max orientations (denoted by bar) or randomly distributed Ro max orientations (denoted by circle), after Stone & Cook (1979b)

The strain figures from vitrinite samples collected around these faults generally show two distinct peaks (Figure 4) . The calculated non-random Ro max orientations show no consistent orientation with respect to the faulting (Stone and Cook, 1979 b) . However, by identifying each of the strain peaks as an individual strain component or event (each of which is designated in Figure 5), a recognisable and consistent pattern emerges . The earliest recognisable strain is a regional event (thin bar, Figure 5) oriented W. This may be equated to the Type I and Type 2 strains, which in this example are coincident because the normal faults have formed subparallel to the regional stress direction as defined by the vitrinite strain figure . The strain component oriented NNE (dashed bar, Figure 5) is restricted to the vicinity of the normal faults and represents Type 3 Ro max orientations associated with strain relaxation sub equent to faulting . The third strain event is oriented E E and has a widespread occurrence. This strain orientation is consistent with a W oriented tress field related to the observed dextral strike- lip faulting. The examples demonstrate that some non-unia xial optical characteristic of vitrinite is representative of strain associated with fault formation . The consistency of the regional strain orientations (Stone and Cook, 1979 b) further indicate that the shape of the indicating surface of vitrinite is a reliable indicator of the lateral strain direction .

REG IONAL EXAM PLE BURR AGOR ANG VALLEY Oriented coal samples collected from four mine in the Burragorang Valley, South-Western Coalfield, N .S. W., were studied to observe the variation of strain measured in vitrinite on a more regional scale (in this study, over approximately 50km 2). The sampling pattern was not a systematic coverage of the area but was biased to the recently The mined and accessible areas (Figure 6). western and northern boundaries of the .mines are • marked by an escarpment.

oo . . I

''°

0 .1

SCA.LIE

FOR

A .. AL LINES

432

REFLECTANC E STRAIN FIGURE

Figure 4

Strain figures derived from samples around the j unction of a strike-slip fault and normal faulting. The axial lines of the strain figure represent section orientations and the reflectance value of the centre is 1.25% reflectance, after Stone and Cook (1979b)

Structure

""' o

som

STRAIN COMPONENTS

Figure 5

Major strain components of separate strain events. Refer to Figure 4 for sample numbers. Thin bar - Type I Ro max orientation component, dashed bar - Type 3 Ro max orientation component, thick bar "strike-slip fault" strain component, after Stone and Cook (1979b)

The Burragorang Y!!lley mines are located on the western side of the "controlling syncline" of the southern Sydney Basin (Wilson and others, 1958), and have strata dipping at approximately 3° in a direction north of east. Small monoclinal flexures are common in this region. Two larger northsout h trend ing normal faults, the Nepean Fault and Oakdale Fault , occur to the east of the mine leases (McLean and Wright, 1975). In the mine area the main structural feature is a zone of northnorth-west trending normal faults ( !Orn throw) with a hinge zone centred about Oakdale Colliery (Figure 6). The sense of throw is different at either end of this zone due to an apparent scissors movement. Associated with the faults are a series of d ykes, which commonly trend east-south-east or are subparallel to the normal faults . On the eastern side of the fault zone, the normally poorly defined monoclinal flexures which occur throughout the leases become more pronounced . In the more southern collieries east of the fault zone, the flexures have steeper gradients but shorter wavelengths, although the overall gradient is similar to the more consistent grades of the northern area. The eastern extent of the flexuring is not well defined but seam levels in boreholes indicate alternate areas with steepening and flattening of the gradient. A regional BMR gravity plan (PJate 5 of Mayne and others, 1974)

85 shows a steepening of the gradient in the vicinity of the eastern portion .of the Burragorang Valley mine leases.

Mine roof conditions Apart from the adverse effects of local features such as deeply incised overlying gullies (Enever and McKay, 1980) and faults , the roof conditions of the mine roadways to the west of the fault zone are generally considered very good . This is also true of those mined areas in Brimstone I and 2 Collieries, which are east of the fault zone. However, to 'the east of the fault zone in Oakdale and attai Bulli Collieries, the mining conditions are atypical of previous Burragorang Valley experience (Nicholls, 1978). Here roof conditions are generally poor with the style of failure indicative of dominant lateral ground stresses (e.g. Aggson, 1979), which preferentially deform northsouth oriented mine roadways (Figure 6). The principal stress direction is thought to be near horizontal and J>riented between east-north-east and east, amrt1iis has been confirmed by an in-situ stress measurement (C.Rawlings, C.S.I.R.O.-pers. The area of poor roof conditions comm.). coincides approximately with the more pronounced monoclinal flexures .

Vitrinite reflectance results The samples studied had Ro max values of strain figure peaks which range between 1.04% and 1.14%. Strain figures commonly have two sets of distinct reflectance peaks, the exception being samples taken adjacent to faults where only one peak is evident. In judging variations in the regional trend , results from these more localised features were not used. An analysis of the strain peaks of the vitrinite indicating surfaces reveals two separate areas (Domains A and B), distinguished by two different strain patterns as summarized by the rose diagrams of the strain components (Figure 7). The main contrast of the strain pattern between the two areas is the orientation of the northeast trending peak. In Domain A, azimuth range from 042° to 072° and in Domain B they range from O16° to 032°. The other main strain peak direction has overlapping orientations in both Domains A and B (A: 123° to 164°, B: 101 ° to 142°), although in Domain B there is a relative anticlockwise rotation of some strain peak orientations. This was also found for the northeast trending strain peaks. The exact nature of the boundary between Domains A and B is unknown, although the two western-most samples in attai Bulli Colliery are approximately 40m apart and show different strain patterns. On a regional scale this is an abrupt change if the sampling is representative.

Relation of conditions

strain,

structure

and

Figure 6

Schematic representation of the main structural elements and the area of difficult mining conditions related to abnormal lateral stress in the Burragorang Valley mines. The sample sites are shown by dots and the extent of mine workings are shown within each colliery boundary

Figure 7

Two separate domains (A & BJ defined by different vitrinite strain patterns. Rose diagrams represent the strain components of the main peaks on vitrinite su:ain figures

roof

Figure 8 is the superposition of plans showing geological structure, the extent of poor roof conditions and the two domains with different v1tnnite strain patterns. The Domain B strain field and the area of poor roof conditions, resulting from a dominant east-west trending lateral stress , are reasonably coincident.

86 Figure 9 shows that there is no apparent increase in strain ellipticity (bireflectance of the strain figure) either with increasing rank or between Domains A and B. Whereas higher rank coal has been linked with greater depths of burial and therefore higher stress / strain conditions (Williams and Wilson, 1976), it does not appear from this study that rank varies with rock mass strain.

•

0-07

z < I-

O·Oo

0 ·05

•

..J

a: iii w z < ..J a.

0·03

•

0·02 0 ·01

w m

Figure 8

Areas of overlap of structural features, vitrinite strain domains, and the region of lateral stress induced roof failure , Burragorang Valley mines

Interpretation and discussion of results The results show that, for the area studied , regional strain patterns are consistent within each tructural domain. Such consistency of results verifies that non-uniaxial behaviour related to strain can be found in coals with a vitrinite reflectance as low as 1.00%. Investigation of lower rank coal has yet to be carried out. In both Domains A and B, the south-east oriented strain is indistinguishable because of the overlap of orientations and might be assumed to be the result of a common stress field orientation. If the two sets of north-east trending strain peaks of Domains A and B were formed by the same stress field, at the same time, similarly oriented peaks should be observed in both domains because vitrinite apparently records strain events until erased by overprinting. Evidence of such overprinted peaks oriented north-east is not provided by the data . It seems unlikely that both strain domains were formed at different times by localised stresses. One explanation might indicate their simultaneous formation by the re-orientation of the regional stress field across some structural feature or boundary, possibly related to movement originating within the basement. The magnitude of the in-situ strain in the rock mass is higher in Domain B, as indicated by its response to mining. An increase of stress field inten ity in Domain B is one explanation. Such local variation in strain magnitude would more likely result from basement movements than regionally applied tectonic stress. Proximity to the escarpment might be an alternative explanation for lower strain of the rock mass in Domain A.

•

0 ·0(

...'? MAXIMUM

Figure 9

= REFLECTANCE

Relationship between maximum reflectance (measured) and the bedding plane bireflectance. Domain A - crosses, Domain B - dots

The roof and floor strata (i.e. non-coal material) in the e two vitrinite strain domains respond differently upon mining. It is difficult to attribute a direct relationship between the measured vitrinite strain figures and the observed physical response of the non-coal materials. The effects of stress acting on both the coal and roof strata might not necessarily be imprinted in either because of the completely different set of physical properties involved. o penetrative fabric has been established in the roof and floor lithologies. The typically poor correlation between coal cleat and joint orientations in the roof strata is an example of this. If any genetic relationship exists between the variation in strain pattern between Domains A and B and the higher in-situ stress in Domain B, it is implicit that residual strain is a main component or source of this stress. An applied regional tectonic stress field would more likely be widespread than restricted to localised areas. One interpretational difficulty encountered with regional strain patterns derived from vitrinite reflectance is in relating them to a particular geological event. Strain recorded in vitrinite around single structures generally appears to be consistent with the formation of that structure. However, on a broader scale, it is at present not known which events are recorded in the vitrinite and which stress pulses remain unrecorded . Hence, the interpretation of the Burragorang Valley vitrinite strain figures is uncertain, except to say that the two areas recognised have different strain histories. These are recorded both by the lateral

87 strain release evident from the roof and floor strata upon mining and by the vitrinite reflectance results. The two strain domains, A and B, are not defined by any change in structural characteristics, such as would be interpreted from borehole data. Structure contours of the Bulli seam in the Burragorang Valley, for example, derived from mine workings and exploration drilling, do not show any significant change in the intensity of monoclinal flexuring. A poor fundamental knowledge of the mechanism of imprinting a penetrative strain on the vitrinite also restricts the full interpretation and understanding of the strain figures. Two suggestions for the mechanism of imprinting have been proposed (Stone and Cook, 1979 a): (i) the anisometric growth of polyaromatic micelles during coalification due to unequal lateral stresses; (ii) a type a mechanical distortion of the vitrinite structure. Of these the first is favoured but neither can be discounted . Some evidence for the type (ii) im1:rinting mechanism might be interpreted in atta1 orth Colliery, where the presence of a number of small low angle reverse fault planes (Im throw) is most likely due to the directional increase in lateral stress in the vicinity of a deeply incised valley (Figure 10). Perpendicular to these reverse faults is a small, high angle normal fault, which probably represents reactivation of a pre-existing joint plane. The effect of valley incision on the redistribution of the stress field has been recorded in outcrop near the study area (McElroy, 1967), as have the stress distributions that might be expected in the area of influence of incised valleys (Pariseau, 1971; Worotnicki, 1969; Enever and others, 1978). In Nattai North Colliery, Enever and McKay ( 1980) have related in-situ stress measurements to the presence of an overlying valley. Beneath the valley, the principal stress magnitude is approximately double relative to the normal stress state. More significantly, the relative values of the in-situ principal stress components ( 6' 1 = 18.5 M Pa, If 2 =4.3MPa, 15 3 =approximately 0) indicate stress differentials more conducive to rock failure than elsewhere. The incised valleys are probably of Tertiary age, therefore being formed after coalification, and appear to be associated with strain figures and strain components (thick lines), that are at variance with the regional pattern, but consistent with the formation of reverse faulting (Figure I0). The dashed strain components in Figure IO are poorly defined, but their orientation suggests they are remnants of the typical strain components of Domain A. Sample 5 in this figure is not related to any known pattern. The above evidence suggests that late stage tectonic episodes may be recorded in vitrinite after the completion of coalification possibly by the mechanical distortion of the vitrinite micellular structure. Much further work needs to be done, to establish the mechanism and conditions of this imprinting.

~----

F1 _ _ ___,_-,o,';,;;;~71-7.-

Figure JO

Vitrinite strain components measured in the vicinity of a reverse fault (FI) of probable post-coa/ification age and a high angle normal fault (F2)

CONCLUSIONS In the weakly deformed area of the southern Sydney Basin, reflectance measurements on oriented vitrinite samples demonstrate that the shape of the non-uniaxial indicating surface can be related to geological strain. Vitrinite is capable of recording more than one strain phase, so that the sequence of strain development around localised structures such as faults can be deduced. To date, significant strain has been measured from the Bulli Coal, with Ro max as low as 1.00%. By analysis of the main lateral strain components from coals measured in Burragorang Valley mines, strain patterns of more regional significance have been measured . Within each structural domain, the strain components have relatively constant orientations, which at least demonstrates that vitrinite may record strain away from areas of increased localised stress such as faulting. Strain figures determined from vitrinite reflectance measurements in the Burragorang Valley region have shown two areas with different strain patterns. One area is coincident with a greater release of strain energy from the rock strata during and after mining. Poor roof conditions ensue as the rock fractures in an effort to achieve a state of equilibrium. On the scale of the mine leases in the Burragorang Valley, the structural style (monoclinal flexuring) is similar, although on a smaller scale the flexuring is slightly better defined in the area with poorer roof conditions. A direct relationship between the variation in vi trinite strain figures and the higher strain state in the rock mass cannot be proved , but at least may provide an empirical indicator of a region with a different strain state. Vitrinite appears to be able to record subt le changes in strain, but the imprinting mechani sm and timing of the strain events in geological tim e are fundamental aspects which need further research. For weakly deformed areas that have limited penetrative structures, this technique of strain determination should be a useful means of aiding structural interpretation , especially in relation to the assessment of mining conditions .

88 Studies of the response of v1trm1te to different structural regimes and conditions are very restricted , and further work will need to be done to adapt the method to routine u e. It would appear to offer advantages due to the relative ease of obtaining and using small coal samples to gain re ults from basic reflectance techniques which are already established in the coal industry. As has been shown in the Burragorang Valley example, reflectance-based strain re ult have depicted an area of difficult mining conditions which is not obvious from structural interpretation, such as would have been gained from normally spaced borehole data.

REFERENCES Aggson, J.R., 1979. Stress-induced failures in mine roof. United States Bureau of Mines Report of Investigations No. 8338, 16 pp Blackwood, R.L., Hargraves, A.J ., and McKay, J ., 1976. Absolute stress investigations in coal at Appin Colliery, New South Wales. Proc. ISRM Symposium - Investigation of Stress in Rock, Advances in Stress Measurement. Institution of Engineers, Aust. ational Conference Publication No. 76 / 4. 17-22. Burger, H . R ., and Hamill, M . . , 1976. Petrofabric stress analysis of the Dry Creek Ridge anticline, Montana. Bull. Geol. Soc. Am., 87, 555-566. Connelly, M.A., 1970. A geological structural assessment at Appin Colliery with reference to roof failure and directional mining. Proc. Australas. Inst. Min. Metal/., 187, 17-26. Cook, A.C., Murchison, D.G. , and Scott, E ., 1972. Optically biaxial anthracitic vitrinites. Fuel, London., 51, 180-184. Enever, J.R., Shepherd, J., and Huntington, J ., 1978. An initial mathematical modelling study of the relationship between underground working conditions and overlying surface topography with reference to the Western Coalfields, N.S.W. CS/RO Aust. Div., Appl. Geomech. Geomechanics of Coal Mining Report No. 5. Enever, J.R ., and McKay, J ., 1980. Stress measurements at Nattai North Colliery and their interpretation in terms of sedimentologica·l and topographic features . CS/ RO Aust. Div. Appl. Geomech. Geomechanics of Coal Mining Report No . 29 Friedman, M., 1972. Residual elastic strain in rocks. Tectonophysics, 15, 297-330. Gale, W .J ., I 980. Palaeostress systems and deformational events in the Lower Hunter Valley. 14th Symposium, Advances in the Study of the Sydney Basin, University of Newcastle, Abstracts, 17-18. Hevia, V., and Virgos , J.M ., 1977. The rank and anistropy of anthracites: the indicating surface of reflectivity in uniaxial and biaxial substances. lour. Microsc. 109, 23-28. Marshall, P., Griffiths, L. , and Lama, R.D., 1980. Occurrence of outbursts at West Cliff Colliery. Symposium on The Occurrence, Prediction and

Control of Outbursts in Coal Mines, Australas. Inst. Min. Metal/. Southern Queensland Branch, 19-39. Mayne, S .J., Nicholas, E., Bigg-Wither, A.L., Rasidi, J .S., and Raine, M.J., 1974. Geology of the Sydney Basin - a review. Bull. Bur. Mineral Resour. Geol. Geophys. Aust., 149. McElroy, C.T., 1967. Valley anticlines in the Blue Mountains, .S. W. 2nd Symposium, Advances in the Study of the Sydney Basin, University of Newcastle, Abstracts, /9-20. McLean, A.J. , and Wright, E.A., 1975. Burragorang Region; in Traves, D.M . and King, D., (eds), Economic Geology of Australia and Papua New Guinea, 2 Coal, Australas. Inst. Min. Metal/. Monograph Series No. 6, 219-224. Nicholls, B., I 978. Pillar extraction on the advance at Oakdale Colliery; in Brawner, C.O., and Dorling, I.P.F., (eds), Stability in Coal Mining, Proc. International Symp. on Stability in Coal Mining. Vancouver, British Columbia, Canada, 182-196. Pariseau, W.G ., 1971. Influence of topography on the pre-mining state of stress. Canadian Rock Mechanics Sy mposium, March 1971, /91-200. Riekels, L.M ., and Baker, D .W ., 1977. The origin of the double maximum pattern of optic axes in quartzite mylonite. lour. Geology 85, /-/4. Roberts, J .C. , I 965. Quartz microfracturing in the orth Crop of the South Wales Coalfield . Geo/. Mag. 102, 59-72. Shepherd , J., and Huntington, J .F ., 1981. Geological fracture mapping in coalfields and the stress fields of the Sydney Basin. Stress and seismicity in Eastern Australia - Geological and Engineering Significance, Australian Geomechanics Society and Geo/. Soc. Aust., Abstracts, 4-5. Stone, I.J ., and Cook, A .C., 1979a. The influence of some tectonic structures upon vitrinite reflectance. lour. Geology, 87, 497-508. Stone, I.J ., and Cook, A.C., 1979b. Vitrinite reflectance as an indicator of stress regimes. Compte. Rendu Ninth Internal. Congr. Carbonif. Strat. Geo/. in press. Varnes, D.J ., and Lee, F.T., 1972. Hypothesis of mobilisation of residual stress in rock. Bull. Geo/. Soc. Am., 83, 2863-2866. Voight, B., 1974. A mechanism for "locking-in" orogenic stress. Amer. lour. Sci., 274, 662-665. Williams, W.A., and Wilson, R.G., 1976. Geological aspects of coal seam roof conditions in the Southern Coalfield of New South Wales. Australas Inst. Min . Metal/. Conference, /1/awarra Branch, 65-76. Wilson, R.G., Wright, E.A., Taylor, B.L., and Probert, D.H., 1958. Review of the geology of the Southern Coalfield, ew South Wales, Proc. Australas. Inst. Min. Metal/. 187, 81-104. Worotnicki, G., I 969. Effect of topography on ground stresses. Proceedings of Rock Mechanics Symposium, University of Sydney.

89 Worotnicki, G., and Denham, D. , 1976. The state of stress in the upper part of the earth's crust in Australia according to measurements in mines and tunnels and from seismic observations. Proc. ISRM Symposium - Investigation of Stress in Rock, Advances in Stress Measurement. Institution of Engineers, Aust. National Conference Publication No. 76/ 4, 71-82.

Worotnicki, G., and Walton, R.J ., 1976. Triaxial *Hollow Inclusion* gauges for the determination of rock stresses in situ. Proc. ISRM Symposium Investigation of Stress in Rock, Advances in Stress Measurement. Institution of Engineers, Aust. National Conference Publication No . 76/ 4, Supplement 1-8..

°'

Computer application in exploration and planning of Hunter Valley No.I Mine P .P. Wootton Senior Geologist, Coal and Allied Operations Pty Ltd Present Address: R. W.Miller & Co. Pty Ltd, 237 King St., NEWCASTLE. N.S. W. , 2300

ABSTRACT Hunter Valley No. I Mine is one of two mine areas allocated to Coal and Allied Industries Ltd from an exploration authorisation 15km west of Singleton. After scout drilling on 2km centres, a total of 16 major coal seams was recognised, made up of varying combinations of 38 seam splits. Although the potential of the area at this stage was very great, a major exploration programme was clearly required for complete assessment of the resources present. It was obvious that a massive amount of bore log and analytical data would be generated by such a programme, and that c,omputerisation should be considered for processing this data and for the drafting of bore and coal sections and contour plans. In this paper, the developments in field logging, data editing and applications to report production, reserve calculations and reassessmen_t of exploration data for mine planning are discussed.

INTRODUCTION The Coal and Allied Hunter Valley Mine areas are located near Warkworth in the Singleton / Muswellbrook Coalfield approximately 100km west of Newcastle (see Figure I). The two mine areas lie within an exploration authorisation area that has been extended twice since field operations commenced in 1974. The mine leases were not defined by area, but by in-situ tonnes of coal in working sections thicker than 0.3m, excluding stone bands greater than 0.1 m thick. For both areas , the eastern , western and northern boundaries were fixed , while the southern boundaries were to be adjusted to contain the appropriate tonnages. For Hunter Valley No. I this figure was 56 million tonnes and for Hunter Valley No.2, 350 million tonnes. The estimation of reserves in these areas and the subsequent positioning of the southern boundaries of these two areas are discussed below. Since exploration commenced in 1974, 375 cored boreholes with total length 32000m and 600 noncored bores of 25000m total length have been drilled. During the early stages of exploration, it was realised that the major geological problems in the Warkworth area would lie in coal seam correlation and in the delineation of consistent working sections within the coal seams over the An indication of the expected mining areas. complexity of the stratigraphy is shown in Figure 2.

GEOLOGICAL APPLICATIONS The method of bore core logging used played a major role in overcoming the problems of correlation. It involved the division of each seam into coal sub-sections or "plies". These were selected and described on the same basis as the

sub-divisions taken for analysis. They were made by selecting coal units which might be recognisable within the seam in a significant number of

Hunter Valle y, New cas t le.

l'te wcaslle Port 150km

LOCATION MAP

C. A A. HUNTER VALLEY Noa. 1 A 2 MINE PROPERTIES.

Sca le · !!!!1111i.aeaaai.,.s;;;;.eff ,_

12 km.

Figure I

Figure 2

Hunter Valley No.I & 2 summary cross section

boreholes. The procedures used in this subdivision were: (a) the proportion of "bright coal" was estimated quantitatively and expressed as a percentage in 10% units; (b) each ply was chosen so that the amount of "bright coal" was significantly different from that in the sub-section immediately above and below; (c) stone bands and units of shaly coal greater than 5mm were excluded . Coal plies selected in this way were rarely less than 150mm thick . Smaller coal sub-divisions were not made because inspections of coal faces in mines or of large diameter bore cores (for example 200mm) have shown the restricted lateral extent of thinner coal sub-sections. Extreme detail in a coal seam log also generally resulted in "not seeing the forest for the trees". This system of Jogging was supported by coal core radiographs on boreholes at 800m spacings . The analytical results from the plies selected verified the sub-section divisions in that quality parameters were generally consistent for the various subsections over the exploration area. Lateral changes in quality of a coal seam were, for the most part , a result of physical variations of the plies within the seam . Examples of this are shown in Figure 3, including changes in the ply thickness and absence of a particular ply or plies through splitting or lensing out.

DAT A BASE DEVELOPME NT Expressing the proportion of "brights" quantitatively is particularly adaptable to computerisation. An example of a bore log printout of such data is shown in Figure 4. Of particular note is the column arrangement of such factors as coal type, non-coal lithology, grainsize, hardness and structures. This enables quick scanning for particular purposes. To the right of the thickness and depth columns, information such as core recovery, stratigraphic names and seam working sections is also shown. Any additional comments are recorded in the space on the far right, including for example, standing water level and other features which the field geologist believes are of a particular significance. Initially, the analytical data was on a separate data base, so that any output requiring both geological and analytical data (except basic lithology) required the setting up of another file or list. Because of this limitation, a subsequent data base was set up by amalgamating part of the ply analytical data with the bore log data base. In doing so the general description section was removed, but ths capacity to store structural data was increased. In many ways this provides a great advantage, as it is now possible to produce sections, blockmaps and tables showing quality and physical information, and to calculate reserves without the additional time-consuming step of obtaining a new listing for each job. A sample of

Figure 3

Hunter Valley No. I & 2 summary seam section

the printout from a draft ver ion is shown in Figure 5, while the coding form for this amalgamated system is shown in Figure 6.

RESERVES Re erve estimates have been made to date using several different sets of criteria, and it is expected that estimates using different criteria will still be required in the future . Some of the criteria used in these estimates are as follows : (a) Reserves for underground mining: Seam thickness ~I.Sm ~0.3m included; Stone bands Maximum percentage of stone bands m workable section 20%; (b) Preliminary reserves for open-cut mining: Seam thickness ~ I.Orn; ~0.6m included ; Stone bands (c) Reserves estimates by Department of Mineral Resources: Open-cut reserves all seams ~0.5m; These must be' worked even if only to be stockpiled until favourable markets exist for the coal concerned; (d) Reserves for definition of lease boundaries in Hunter Valley os. 1 and 2: Seam thickness e!' 0.3 m; ~ 0. 1m included; Stone bands (e) Reserves for mine planning and feasibility studies: Criteria determined by the planning engineer, considering working thickness, parting thickness and nature, roof and floor lithologies and coal washing yields;

Reserves for operational purposes: Criteria adjusted as mineability is determined in the pit, based upon experience with the physical conditions encountered. In Hunter Valley No. I mine, there are nine working sections, while in Hunter Valley No.2 there are 35. Consequently, the task of calculating the reserves for the two areas was immense. However, when the actual lease areas were to be determined from in-situ tonnages of coal, the calculation took on great importance at this early stage of mine planning. Failure to solve this problem would most likely have delayed the granting of the mining lease, as well as mine planning and development. U e of the computer to set up a number of areas, estimated to have approximately the required tonnage, enabled the calculation of reserve and definition of the final areas to be obtained within hours.

(f)

COAL SEAM CORRELATION The features generally used in formation correlation from bore cores were as follows :brightness profiles; "marker beds "· general formation types or cycles over thicker parts of the stratigraphic section. A summary of stratigraphic variations in the area is shown in Figure 7. This indicates that there are several cases where the use of one or two of these features would produce either erroneous or non•• unique solutions.

.,,m .. HA Otll l N

...

AN O A~ ER.,.A I ' I S ( AH AM

CO L LIERIE S LT O.

"' AFUC. 111 0R TH PR OJE CT

<co-.., r. ,

AN &t q

~ D"

--------------------------------------------------- ------------- -------------------------------------------------------------- THI CI< . r'I E~ TH REC • ST 0 AT. GRAI N

CO LvUK BR[l,j,HJ •

Ll T"OL UGY

HARO ,

S TA Uc T .

s n::i:. u

OESCRI PT ION OF

. ... ... .... . ... . ... ... ..... ... .. ..... . ... .. .

,.,

T O c:J Ac;E

r l')" ""'° NTCi

---------------------------------------------------------------------------------------------·--·---------------------------,u, 8~ IGHT CO aL o .3 q o 77 SHALE B ANO S 7 0 1 5 0~ ft4 FRO"'! B ASE 7 2. 2 5 0 L "'A l I

MUOS TON E d ;;jO • N SQ I Gt-H COAL MVO <"; lO E awO • N COAL dAI G1-1T 14Ut ,S TO,•E GA EY - riAh s .. I GH r CO • L B.:tlGHT CO• L MU1,S TO E Gol[Y 81.flGHr COa L MW ,S TONE GQ[Y ::- 11 rSHL . Gi.f[Y G,.El S. A...o S T. ~ I L ( SHL . b .. EY ::.1 ,,r;c 1 rE s wo w:,,, ':> 11 TSrtL . G,. E Y MU11t;TO E b .. EY cv .. L B ~I .i i◄ T d Wl b Hf CO.. L MU, 1S f0 1•E GQ [Y di.tIGHT CvaL Jrolv , , ST 0 "4E G_, EY di.. l GHT CO.. L "'IJ•1~ T0 ,,1E li"'EY dw l GHI CO•L 114U,1'.:lo T.0 1-.iE 1J of EY B~IGHT I..Oi.L '.:lo.t, r,.nS T. u .. E Y- r,A i-.i fll!UnS T0 1,.E GA EY-HAli 8id GHT CO• L i:t'l l GH T CO• L MUI, S TONE G.,.EY d ~ ICii H T l.OA L B'l l liH T CO aL MU11 Ci TONE i.;,~EY '::I I L TS.HL • i.;,"'EY - .._ RN 6~IG1-H CO a L 5 11 JSHL. Gi;;ET SANn ST • G_,ET SH• L E d WI GHT CO •L SANOST • GA T- • HTE G_.EY ':> Ar.iQST. SH aLE CO a L

"4VO rU, AO

su , MOO 1-t ARO

ros o<

PELLE T O I n

MOO HARO

S1-i ALE SANO 8 5 "' "' FA0\11 B ASE 5 "4Jl4 COAL B AND i E &A BASE

AT B ASE

3u•

au, T O•

M "4[ U- CR S

LI n<

MOO HA;I Q M(JQ HARO hlOQ HAN O MUO ,U k Q

CO AL B ANOS hll n • RASE S ! OERITIC I N TOP 10 0 M"'11 .30 "1~ S I OERITE c ANO 12Q WM F qQ,114 B

HA l'i Q

MOO HAHQ SF T- HA l'f O

°' °' ""' n<

bO l"M COAL Y B AUO 6 0 "4liill F DQliill TOP CARS . AT ~ • SE

3 0, ru ,

SHALE B ANOS Mto .

TO B ASE

MVQ t-U," 0 Si.t ALE B AND N[.t, A FIASE C') AL Y AT 9 ASE

SF f-H AkO

TU<

n•

Muu l'i AHQ

su ,

"""

T U<

COAL PENNY 8 AW) S NEA R TOP , ' B ASE

HA"D

F i ri E - ME11 MuO HARO MOO rt AMO

MUDDY AT TOP• CO PENMY R Ar~r) S AT T PE LIT IC AT B ASE

MuQ HA RO

SHALE B AND 35"'1"! FR O"! a a SE CARS • AT TOP • CO P(NNYSA NO S AT ~ I

1 0, s u,

n< 10 , so,

SF T-H Ai'tQ

l S"!M CO AL B AND lSMM F ROM TOP Mt NOR CO AL PE N H R ANOS N(Aq BAS E

LT RU <

""'

MI NOR co,L B ANOS IN TOP t. Qli(M S tOE R ITrc B ASE S ANDY TOP COAL Y NEA R S ASE ?S ANO l 0 '4fl! FROM TOP

F' (NE -ME O H A~O SOFT ME 0 1UM FI E

Gi.f E T

d .. I GtH

MOO H ARO

H ARO ~00 HA'10 JO INT ED

s-. I NTE RA EOS SS/ SH S0 /50

MOO SOFT

SUBVERT JO INT l M F ROM RASE VSOFT AT BASE CARS. BASE .

R P,P lol,t,P

Q. O't 5 0 .21 0 0 . 025 0 , 66 5 X 0 . 080 0 . 76 5 0 - 90 5 0 .1 10 o . i no 0 , 9 05 o . no 2 - 660 0 . 610 0 . 25 0 1 . sno o . 20 0 X 0 . 420 1 .i 10 o . ots 0 . Q75 t> -1 20 o . 22s o . 0 1s O ol~O Q. Qll O o . eoo 0 . 57 0 O.t 2 0 0 .t 50 0 .310 0 .11 0 o. 160 o . 320 0 -1 40 o . 71.0 0 . 0'50

SH ALE B ANOS 2 0 • 120'4 ... FR0 "4 TOP

72 . 295 72 . 505 72 . ~ 3 0 73 .1 95 73.2 7-; 7 4 . 040 74 . 9 4 5 75 . 07~ 75 .1 7 ..

GfOL.: 'II QPI(

•=

"""

...

l .,. U

3 . J, ~ 3 . '1 Ii

L WAt L WA t L ""'& t L "'Al L "'At L .,.At

76 . MO 76 . 8 1 1'1

7q , t1 1 0 PO . nen PO . 33'1 ft t. "5n

P2 . 0 3 0 -, 2.451"

83 . 560 -, 3 . -; 7-.

8 3 . 650 •J . 77"

U"'P t IJ ¥P1 U"'Pl U"'P1 IIVPt

e 3 . qq<

U1.1p1

I' ll . 0 1 0 ~ll . 19f"t /1 4. 2JO

ll l#P l I J1r,11p 1 ll "'P t

AS.nJo

u up 1

11s.6on 111,'5 . 7 2 '1 p5 . ,:p n

R6. 180 8 6 . 2 9 '1 ft6 . tt5n 86. 77 0 ft6 . 9 l n 1' 7 . 6 2 n ,- 7. 67'1

0,580 0 , 6 20

50•

L MA t

L""'Al

L PFt LPFt LP F 1 L PF 1 L PF t

G~ l)L . = WOQK . :::

...

RP• GF"OL . = wo q K . :::

3 . 11 nr, 5 . nr, n

77 1 . '1 11:in 1 .'1 5 0

ri r EPpu=· o FQ O.. Afll.2 5 u 10111 17q T n 2 7/11/79

M . 25 0 82 Rft . 8 7 0 -. 9 . 471"1

0 , 600 0 . 2 10 1 .1 20 4 . oeo

qo . 1300 oct . 8 8 f'I 9 7

o . 71:10 X 0 .1 3 5

os . 620 q5 . 755

R9.68 0

LPR n.l t O GE'OL . = FHT NOV 7 g SWL 11 . 01111

UV T

F MT NnV 79

Figure 4 ODH E AST I UC NORTH ING

22 4244 000 I ~b06 48 000

COLLAR RL TOT AL DEP TH ·

STANO l NC WA TER LEVEL : CAS INQ :

.....

3 39 . 59 02 . 100

DATE COMNE.NCED . DATE COMP L E TE D

• NO CONT

227

ST PLV

R D I NSIT U <A I R DRIED DASIS I •PT I H. ll()L ASH SUL CSN S E

0 2/02 / 8 2 0 9 /02/02 CORE St ZE · LOCCED BY .

DAT A Ol'ISE PREP ARED DY EAR TH SC l ENCE COMPUTER SERV I CES

NO P WOOTl ON

•• ................................................................................. ................................................... 13 0

ROCK TY PE

CRAIN COL OUR / MEC HANI CAL STATE TH I CK . DR I CHTNES 2 ND. L IT HARO . OISCON T .

SANOST SHALE SH / CARD I C ROCK S I L TSH COAL / ST S I L TSH SANOST . SHAL E SANDST SH / CARD SANDST . COAL / S T SAND ST SIL TSH COAL SHALE SANDST SHALE SH / CAA B SHALE COAL I1UDST COAL SIL TSH COAL NUDST COAL NUDS T COAL COAL NU DST COAL COAL 11\JOST COAL NVDS T SANDST COAL COAL SH/C ARB

CREY - BRN MED I UM SAND Y CREY CRE Y LT SAND Y CRE Y SANDY NEDJ UM

HO- SFT SFT-HO HARD HARO HAR D

CRE Y CRE Y CRE Y CRE Y

ME DIUM

CRE Y

ME DI UN

HAR D

GRE Y GRE Y 3 CREY CREY CREY

MEOI UN SANDY

HAR D HO - SF T

f1EDI Uf1 SANDY

SFT-HO MARO SF T-HD

WEATHER . INDURATE

HAR O HARD SOFT HARO

JOINTED

CREY

• •

SFT - HD SFT- HD 10 SHAL.E Y

CREY 20 SHALE Y CRE Y-BRN

• • • •• • ••

JOINTED

2 0 SHALEY

BL ACK

SFT - HD HARD

•o

CREY- BRN

SHALE Y 30 CRE Y- BR N 30 SHALEY 50 CREY- BR N b0 CREY COARSE WHITE

HARD

SOFT

SOFT SFT- HD HARD J O I NTED JOI NTED

DEPTH

REC

9 . 450 2 2 . 4 20 95 AR 10 . 7 10 33 . 130 99 0 . 770 33 . 9 00 HV A 0 . 500 3 4 . 4 00 HV C 1. 2 70 35 . 670 0 . 2 7 0 3 5 . 94 0 4 . 260 4 0 . 200 2 . 350 4 2 . 550 3 . 850 4b . 4 00 1. 130 4 7 . 530 0 . 560 4 8 . 0 9 0 b00 7 . 510 0 . 250 8 50 1. 370 57 . 2 2 0 1. 3 8 0 58 . 6 00 0 . 19 0 58 790 0 940 59 730 3 . 990 b3. 720 2 . 480 66 200 0 . 100 66 . 300 4 . 6 70 70. 970 0 . 190 7 1. 160 84 HV E 0 . 4 00 7 1. 560 HV F 0 . 390 71 9 4 0 HV 0 0 . 890 72 830 HY M 0 . 2 60 73 090 9 2 l'1V I I 0 . 030 73. 120 HV J I 0 . 060 73 . 18 0 HY J 0 . 0 3 0 73 . 210 HV J 0 . 21 0 73. 420 MV K 0 . 1 BO 73 . b00 11\1 K I 0 . 090 7 3 . b90 MV L 0 . 330 74 . 0 2 0 HY 11 o . :270 7 4 . :290 HV 11 0 . 020 7 4 . 3 10 HV N 0 . 180 7 4 . 490 HV 0 1. 050 7 5 . 54 0 HV P 0 . bOO 7 6 . 14 0 HV P 0 . 340 7 6 480 HV Q 1 0 . 290 76 . 7 7 0 HV S I 0 . 060 76 . 8 3 0 HV T I

CUN FI b 0 CUN F I 40 MSX ASH CSN NS X ASH S E

"" ··

1. 46 1. 88 I 2. 02 1. 46 1. 8 4 1. 84 1. 84 I. 44 I. 33 1. 9 1 I. 33 I. 3 0 I. 52 1. 33 2 . 2b 2. 26 1. 40 1. 3 0 1. 6 5

Figu re 5

29 . 5 b3 b 25. 3 75. 3 30. 3 66. 7 66 . 7 66. 7 27. 4 lb. 7 74 . 3 15 . 3 13 . :2 73. 2 14. 9 9 1. 4 9 1. 5 17 . 9 II. I B

...

83 27 0 9 2b 7 BI 20 3 2 28 I 6b 21 B 0 0 0 78 2 4 0 94

14 . 4

0 94 94

12 9 II 1

0 89 10 8 99 17 9 9 8 10 . b 23 29 . 2

95 PROJECT /CO LLIERY

Grllf)hic

SURVEY

Punc::h

Figure 6

:::::::::_

ANALYSIS

ICORRELAl' N CHECKED I

,.,. ........•. •' ··• DATE

Bore log coding form LEGEl'!..Q

~ 5-15

~ E ROSION

~ THINNING

t,RADUAL

W EDGE /SPLIT

£ DETERIORATION

~ CONGL OMERAT

SANDSTONE

~ M UDSTONE

□ STONE/COAL

I] COAL : DULL ■ COAL

BRIG HT

C. g A. OPERATIONS . HUNTER VALL E Y N'l1+2 SEAM VARIATI ONS /I

Figure 7

Hunter Valley No.I & 2 seam variations

, ff O

96 While the correlatio n achieved in the Hunter Valley mine a reas has been reasonable, it has evolved and bee n modified si nce 1974 using one or two methods (particularly brig htness logging). It is likely that a more accurate correlation would have been ac hieved more rapi dl y if a ll three features had been ext racted from the bore log data at the one time, an d presented over the com plete ection penetrated by each borehole. Compilation of detailed gra ph ic logs of coal units and interseam secliments, where the three main lithological types , sandstone / conglomerate , shale/ mudstone and coal are plotted at different cales, based on the compaction ratios of the various units, is relatively simple with computer technology. Such logs enable a n "as deposited" sedime nta ry model to be built up for particu lar "instant s" of geologic time, and hence allow improvements in bore-to-bore correlation. Even if th e compaction ratios assumed are not un iversally accepted, the concept ma y still provide the necessa ry stratigraphic data in a ph ysica ll y ma nagea ble format. An indication of zo nes of poss ible instability due to differential compaction would also be given by such a plot.

coal reserves, but a lso on the economics of the mining ope rations.

ACKNOWLEDGEMENTS The author is indebted to Coal and Allied Industries Limited for permission to publish this paper and to Earth Science Computer Services Pty. Ltd., particula rl y Dr Bob Johnson and Mr Paul Wallin, for ass ista nce and advice.

CONCLUSIONS The initial decision to establish a computer data base for both bore logs and coal analyses was made on the basis that the amount of data likely to be generated from such a large drilling programme would be a lmost impossible to utilise if plans and sections were to be produced in a reasonable period of time. However, it also appears that, on a cost basis, and because of the extra flexibility that arises from a well-ordered data base, the application of such techniques to conventional sized colliery holdings (e.g. a collie ry with 20 boreholes and 50 seam sections), for routine reserve up-dating arrd qui ck output of data for mine planning is strongly recommended . The bas ic borelog report itself is far cheaper when processed through such a system. Data punching, prepara tion a nd printout for a typical borelog and analysis report costs a round $30.00. This co mpares most favourably with conventional borelog report preparation which would involve at least four hours typing time of th e geological log alone. Thus, for a cheaper input cost to that of standard procedures , a far more powerful system, ena bling quicker and cheaper output, is provided. In applying computer-based data processing techniques to coal ex ploration and mining, the most important factors are that there is a clear understanding of the objectives, so that the data base can better enable such objecti ves to be achieved , and tha t the system has flexibility. The geologist is only th e first person to use th e system . Mining and ex plorati on technology is alwa ys subject to cha nge, as are marke ting situa tio ns. If the system is to fully utilised , it should be able to handle the differing requirements of the va rious pecialists whose res ponsi bility it is to prove , produce and sell the coal. In the long term, the data produced should be able to be used not just to provide info rma ti o n o n geological features or

Application of geology 1n longwall m1n1nga case history of longwall 202 in the Bulli seam at South Bulli Colliery Bulent Agrali The Bellambi Coal Company Limited, P.O.Box 17, CORR/MAL, N.S. W. 2/58

ABSTRACT Longwall mining provides sixty to seventy percent of the total production at South Bulli Colliery. Underground geological work is therefore mostly concentrated around projected longwall blocks. Very detailed mapping and close-spaced measurements carried out in development workings, together with an adequate sampling and subsequent analysis and testing of the coal seam, form the basis of the geological advice in longwall planning. All variables, such as roof lithology, jointing, seam thickness, the occurrence, size and extent of stone-rolls, the extent of igneous intrusions and of associated cinder, and all coal quality and washability data are thoroughly investigated. The variation patterns of these features are used in determining the most economic shearing height, and also to predict mining conditions and likely quantitative and qualitative variations in the run-of-mine product. Together with similar predictions formulated for other longwall blocks and - though with much less precision - continuous miner panels, this geological work may allow better short and longterm mine planning by setting variable advance rates for different units and for different subsections of longwall blocks with the ultimate aim of setting reliable quality control standards. Inspections of the face are carried out at regular intervals during the life of the longwall block , and all observations recorded. Initial predictions are compared to actual conditions, discrepancies are analysed and the cause of unforeseen variations in the dilution ratio are assessed as being either geological or operational. This analysis may eventually lead to a modification of the basic prediction method . The case history of the longwall 202 block dealt with in this paper illustrates the step-by-step approach used in geological forecasts , from the early stages of "district planning" to the final adoption of an extraction pattern .

INTRODUCTION In a colliery where the saleable product is expected to conform to stringent export specifications , the role of the mine geologist in quality control overlaps, and even predominates, his normal duties related to strata control problems. In the special case of the South Bulli Colliery at Bellambi, .S. W. , where the final saleable product is a blend , in variable ratios, of coal from two seams (Bulli and Balgownie) with different quality and very different mining characteristics, the solution of various strata control problems is often presented as a multi-faceted answer to the multifaceted question: 'How would this problem affect : (a) the composition of the R.O . M. product? (b) washed coal characteristics? (c) the performance of the coal processing plant? . and , how and to what extent could it be remedied by: (a) underground support and mining techniques and / or (b) appropriate modifications to the mine layout and / or production targets for various units, and / or (c) coal processing by alteration of the was hing medium density and / or blend ratios, separate stockpiling etc.'

To view the role of the geologist under such circumstances it is essential to emphasize that , unlike the Balgownie seam, the thickness and quality variations of the Bulli seam ma y occur over very short distances, even within a single longwall block, despite the indications from widel y spaced boreholes of regular directional trends. Longwalls are high production units, and the impact of their mining and coal quality characteristics is immediately felt in the coal processing plant , particularly when this impact is negative. It is therefore essential that both the mining and the coal quality characteristics of a projected longwall block be known in minute detail before the start of production from the unit. Geological advice may provide a sound basis for setting the layout for a new mining district (early planning stage), for determining the most economic dimensions of individual longwall blocks (late planning stage) and in suggesting desirable rates of advance for the whole or different parts of a given block . Eventually, all this may help in assigning target production rates for individual mines and individual longwall or continuous miner units, and in working out a flexible budget for t he coal processing plant. The compilation and interpretation of geological

---,98 data, and the presentation of geological advice and suggestions are carried out, as a matter of routine, in the form of a small-scale feasibility study. The case history of the "Lo ngwall 202 Block" dealt with in this paper is a typical illustration of the role played by a mine geologist at vario us stages of mine planning a nd , later, as the extraction of the block proceeds.

LONGWALL 202 The geographic location of this block is shown in Figure I. The dimensions of the portion destined for extraction by longwall mining were I 31 m (face) by I 070m. Extraction, carried out with a fully integrated W-SD unit, started on 24 August, I978 , and was completed on 20 July, I979, in 34 Initially, before the development work weeks. started , it had been assumed that the block would yield 482000t of raw coal, while 61000t of various floor and roo f rocks would be, willingly or unwillingly, mined and conveyed with the coal. With a predicted laboratory recovery of 89.8% for coal at 1.50 R.D. , this would mea n a 79.7 % washery yield . However, when the development work was completed , it was re-calculated that if the whole block was mined as planned , the average washery yield would fall to 68. 1% for the sa me washing medium density, with local slumps to as low as 30%. The case history that follows ex plains the reason for the discrepancy between the initial predictions and the actual situation, and how the problem was eventually solved.

EARLY DISTRICT PLANNING A preliminary geological assessment was made, in August, 1974, of two new d istricts, named R and S, sit ua ted to the west of th e Cataract Reservoir, with regard to longwall extraction (Figure I) . At this early stage of mine planning, geological advice was bound to be sketchy, as the districts, with a total surface area of about 3km 2, were covered by a n inadeq uate number of bore holes drilled on an irreg ular pattern, a nd the intended development was literally, to form a deep westerly incursion into a virgin coal domain . Predictions of the possible occurrence and / or frequency and / or location of dykes, faults and stone-rolls , and the likely pattern of sandstone channels in the roof proved to be basically inaccurate, especially for the S-District of which the longwall 202 eventually became the first block to be extracted . Based on a study of airphoto lineaments and surface joints (Figure 2), a suggestion was made for a district layout with the faceli ne oriented 0I0"-190" . The 100"-280" direction, considered to be the best mining direction , was reserved for gate roadways. This prop osa l, ado pted in 1974, was later dropped in favour of a new layo ut a t right angles to the initial one, with longwall faces running 100° -280° , as practiced in the neighbouring R-District. Economic rather than geological considerations, such as the additional dn l'lo pmc nt time needed to prepare a new district

Figure I

Location p lan of R- and S-Districts in South Bulli Colliery Holding

JOINTS IN IMMEDI ATE ROCF

SURFACE LI NE ATI ONS

"'°

,.,. f.,,.,

~--..(

,,. \.

!:t §:;

'-" --t ~

I ~

• ""

Figure 2

H.l.L

·RCl[

JCl •·•

Rose-diagrams of airphoto lineations above LW 202 Block, and of roof joints in development workings (% in length)

with a different orientation, determined this final choice, which provided quite satisfactory, if not perfect , roof conditions. A potential hazard map (Figure 3) had been prepared at this time for both districts, based on a combination of trace density per unit area and length of li neations per unit area. This indicated two areas with potentially difficult roof conditions, both of which were interpreted as zones of intensified jointing, possi bly intruded, dying out in a north-easterly direction . Actually, when the block was fully developed, the southern "potentia l trouble patch" proved to correspond to one of the worst areas with regard to roof conditions, while the northern patch remained unrelated to any geological feature in the underground workings. For both districts, no dykes or faults, other than likely or possible extensions of th ose known in old workings, could be predicted from this airphoto su rvey.

0 [

Jor

101 1

- - - - - • II

.,......

Figure 3

'11_1, Tl)tt

,, . .....,., p,- .,.,o,.o

Potentiai hazard map of L W 202 Block, also showing airphot o lineatiuns

LATE DISTRICT PLANNING After one longwall block had been extracted, and another partly developed in R-District, a definitive layout was worked out for S-District in early 1977, and the geologist was requested to submit mining and quality predictions for individual longwall blocks within the framework of a 5 year development plan. A fresh study of airphotographs was carried out at this stage a some new dyke directions (052" to 065° ), hitherto unknown in old workings situated east of the Cataract Reservoir, were observed in R-District. The fact that the major dykes intersected in that district had no surface expression, while a number of strong surface lineaments proved to correspond to minor zones of intensified jointing in the workings, rendered hazardous the task of predicting any igneous intrusion. Moreover, a cursory survey carried out with a portable magnetometer indicated that, while some of the known dykes caused a detectable anomaly, a number of other anomalies observed could not be related to any geological features known in the workings. A summary of all predictions formulated at this stage for the longwall 202 block is given below, together with the actual values calculated after the development work was completed: Seam thickness (m) % Raw Ash Yield at l.50R .D. % Ash at l.50R .D. Crucible Swelling No.

% sandstone roof area

% stone-rolls in seam (by weight) % igneous rock in seam (by weight) % cinder in seam (by weight) trend for ash increase trend for yield increase trend for C.S.N. increase

As predicted 2.4 13- 14 89.8 11.3 4-4 ½

Actual 2.62 14.7 89.0 10.75 4

4 5.7 1.0 1.5

12.5 5.2 1.2 5.1

Northwards Northwards Northwards

Northwards Southwards Southwards

It thus appears that borehole data was, in the case of this particular longwall block, sufficient to assess the average quality, but the trends for variation of individual characteristics were revealed to be unpredictable. When the above forecasts were formulated , the SE corner of the block was merely 350m away from the advanced workings while this distance was 1000m for the NE corner.

DEVELOPMENT OF THE BLOCK Development work started in SI and S2 panels, destined to serve, respectively, as tailgate and

maingate roadways of the future longwall , in August-September, I 977, and the work was completed by June, 1978. Meanwhile, a d yke, 0.5m thick , had been crossed in the outbye RDistrict, with a 075° strike, and it wa anticipated that the same feature could be intersected in both SI and S2 panels somewhere inbye of cut-through 7 or 8. This was not a disturbing anticipation since, in budgeting, a contingency of 0.8% to 2.5% in weight of igneous rock and associated cinder is regularly made, in line with past experience, for all planned development areas. A high frequency of NE-SW striking joints, alternatively dipping to NW or SE, was encountered a short distance outbye of cutthrough o.4 of S2 panel (Figure 6), which wa s driven in advance of SI panel , and kept intensifying further inbye. Two major roof falls occurred in the stores road (Figure 7, northern heading). Going back to the airphoto lineament map and potential hazard map (Figure 3) became a necessity, and the continuous lineament crossing the block between cut-through o.9 of S2 panel and the intersection C 11 (southern heading) of SI panel, with a strike of 047°, was considered as the surface ex pression of a possible dyke or fault extending to the edge of the Cataract Reservoir , ,1 a northeasterly direction. Such a dyk e, it wa~ thought , could ha ve a southeasterly dip of 78° to 83° and have an influence on roof conditions where the S2 headings were being developed . Some field work followed, revealing no dyke or fault , though it is extremely difficult to pick any surface evidence of such a feature in the Hawkesbury Sandstone, which has an average thickness of 150m to 160m over this particular area. Moreover, the company had recently had the experience of a small dyke intersected in a ventilation shaft stopping sharp at the base of the Scarborough Sandstone. The view was therefore adopted that this would be a small clay dyke not reaching the surface. A zone was reached , 165m inbye of the point where the first joints were encountered, where the coal was fully cindered, with many mall dykes in the coal as well as sills of all sizes and shape . Sixty five metres further inbye, a 1.15m dyke was intersected and, after a 23m interval, a smaller dyke of only 0.3 5m in thickness. Beyond the second dyke, roof and floor conditions were back to normal and there was no cinder (Figures 4 and 7). The same dykes were also intersected in SI panel, inbye of cut-through o.8, with thicknesses of 0.6m and 0. 7m respectively, and a much narrower cinder zone (50m as opposed to 92m in S2 panel) . The actual strike of both dykes was 054° 30' so that their relation to the 047° surface lineament could not be proved. On the other hand , the 065° striking dyke in RDistrict appeared to be an off-shoot , or satellite, of the main dyke system rather than being its easterly continuation. This fact was proved much later, when development work was completed in RDistrict.

The two dykes, undetected from surface· lineaments or by a cursory magnetometric survey, thus occupied the northern margin of a 225m wide zone of intensified jointing, which was apparently widening in a south-westerly direction. Another important fact was that the first dyke was slightly faulted, with downthrow to the north varying between 0.45m and I.Orn. Further north, development work was pushed ahead without any problems, except for a 0.1 m thick dyke striking 280° , in the northern end of SI panel, which determined the -positioning of the longwall face headi ng.

Figure 4

Roof lithology. Stone-rolls. Dykes and cinder areas

PREDICTION OF MINING AND COAL QUALITY CHARACTERISTICS Plotting During development work, the duty of a mine geologist is, as a matter of routine, to observe, descri be and measure all geological features, and to plot this data on separate plans at an appropiate scale (1:2000 in South Bulli Colliery). Variations in seam thickness, roof lithology, roof and floor conditions, igneous intrusions, cinder zones, stone-rolls, roof joints, leaners (low-angle joints), cleats and the occurrence and continuity of in-seam bands had thus been measured , surveyed and plotted in the course of weekly or bi-weekly inspections in SI and S2 panels.

Figure 5

--- -

........

- ~ i . . i , . . . . , . , l ~.,GN

Figure 6

Interpretation When the develo pment work was completed, the usual procedure for any longwall block was carried out for this block as well . This could be summarized as "interpretation of scattered data", by means of which the areal extent of isolated features is investigated and correlations attempted. Figures 4, 5 and 6 illustrate the work performed at this stage. The initial plan of "spot seam thicknesses" is transformed in an isopach plan (Figure 5), while likely continuity and correlation of stone-rolls, dykes, sandstone channels (Figure 4) and joint zones (Figure 6) are tentatively established. The plan on which roof conditions are plotted (Figure 7) is usually prepared as an overlay to be used in connection with all other plans in order to understand the causes of falls and heavy flakings . The isopach plan serves as a basis for calculating the va riations of average seam thickness along the longwall face and, eventually, in determining the most economic height of extraction that would allow the highest possible ratio (rather than the highest possi ble tonnage) of coal in the R.O.M . product (Figure I I, top) . A rose-diagram of all measured joints in the immediate roof rocks is also prepared at this stage (Figure 2), and a general assessment made of particular joint sets likely to ha ve an influence on roof conditions along the face line.

Coal quality plans For this Jongwall block, analysis data from 15 strip samples was used in preparing various coal quality maps. Samples were usually taken every 150 metres, from alternate headings.

Seam isopachs

Figure 7

Major jointing in immediate roof rocks (Zones of intensified jointing represented by their approximate axes; width in metres when >3m)

Roof conditions in first workings

!so-recovery (Figure 8), iso-ash (Figure 9) and isoswell (Figure 10) plans for washed coal at 1.50 R.D. only are included in this paper. Actually an iso-ash plan for raw coal and an iso-recovery plan for 10.6% ash were also part of the original report. Variation ranges for the three characteristics are substantial (see below), and such ranges are usually expected to occur over an area as large as a colliery holding rather than a single Jongwall block: Yield % 74.2 to 96.0 Ash % 9.7 to 12.4 C.S. umber 2 to 5

Prediction of roof behaviour (dilution from roof rocks) Roof falls along the face may have a geological or an operational cause. Geological causes can be predicted by the geologist, whereas operational causes cannot be predicted in time and space. Long delays occasioned by mechanical and ( or

101

For budgeting purposes, however, the calculated volume is increased by a given percentage, which represents the "accumulated discrepancy" between predicted and actual spillage rates for the bulk of The method is previous longwall blocks. empirical, but has been found to work well.

electrical breakdowns, strikes and the like, are grouped as potential operational causes, favouring bed separation and build-up of high pressures along the face line in front of the chock . An operational delay occurring while a geologically sensitive zone is being crossed is likely to cause a roof deterioration out of all proportion to the geological cause, which may extend all through the face. Recovery in such cases is often difficult. and such features escape any logical prediction . Apart from these particular situations, norms have been established in South Bulli Colliery, after many years and the extraction of a great number of longwall blocks, for various "fall triggers", which may be stone-rolls, leaners, bedding planes, dykes, lithological transition zones, areas of critical joint strike and frequency, distance between immediate roof horizons and the base of the overlying sandstone, etc. The amount of dilution from roof rocks likely to occur at any given position of the face line is estimated after the nature and the number of the so-called "fall triggers" anticipated along the face are evaluated. The height and width of each single potential fall is calculated separately according to established norms and taking into account the angle between the face line and the strike of the feature concerned. The resulting height of roof rocks, in centimetres, represents the "predicted dilution" for this particular position (Figure 11).

' -

--i

~ -♦

Crucible Swelling Number (Washed coal at 1.50 R.D.) --

...

-.. r----

- --

r~"

---- ----

- - --- r •• I

==-:::.

,._ lilllll

n :::~. = m,11111

SUGGESTED l.fEIGHlS OF E ~lR ACTIO"

j 1,_ :

I ""

hrTgt!t ol elrt1"0ctn:n

I JO

s,,,.._

• 11111111 ITTllnii

111111 """'"

DILUTION FRCN 'ITICIF'AlEO qoo,:q()Q(S

~ 10

- -~ '11111111 11111111 1'1'1111

20 • jN TIQPAT[O DIL U TION FROM SfCN:-RCl.LS & ~ LOJAROCK$

"""'

lllllllu " '"11'''111111'

11/>>... / )-.

;;~ y;. ,,, ~ ~

>IIEIWE

IS AM TMICKNESS f LONG FACE LINE ;

c knei s

IL' I='=' ts. •

1--i

•• l~'bmr 11!illlll 111111111 "'1111-

,N TIOPAU:0

/ /

DILUTI~ FRi)M

IGNEOUSRCOCS

l{ N~

~·

I :· \ ; \

'\,

Figure JO

.....__ _..,

- 1=

Ash in washed coal (Cum . floats at 1.5 R. D.)

-- t -

30 11,7

s1.: ........

,,. ...

-~-~-··_._-✓~--

\J:.._:.:·~--:~.. :·~~ -_

Figure 9

----== f=L

( _·,,.

·,

- . --..... _ -

...

' I

---- ----- I- -;\.

~g_

~:s

-~L

Laboratory yield of seam at 1.50 R. D. (Bands included)

Figure 8

101()

1100 : LW F1Nlsli

Figure I I

-~ ,po

,i,o

•NO

AS$0Cll>~ D tlNCJE'R;

/~ ½ ~~ >CK<!>< X'>t =

en~.

50 ~ ~I<> ~'C? c e - tstott tine l ~tre-sl -

,po

100

LW slART

310

210

Variations of the average seam thickness. Anticipated variations of likely maximum dilution from various sources.

102

Linearisation of data. Figure 11 summarises the expected average dilution from various sources along the longwall face, for any position of the face, in relation to the proposed height of extraction. Coal quality data from Figures 8 to 10 was also put in a linear form (Figure 12, and the dashed line of Figure 13), in order to predict the quality characteristics of the coal to be extracted at any position of the face line. Prediction of the composition of the R. O.M. product. The thickness of the coal seam proper, for which quality characteristics have been established, at any given position of the face line, was obtained by extracting the predicted thickness of floor rocks plus stone-rolls plus igneous rocks and cinder from the proposed shearing height for the face position considered. Then, by adopting appropriate relative densities (for longwall 202, these were 1.38 for coal, 2.49 for roof rocks, 2.48 for stonerolls, 2.47 for floor rocks, 2.5 for dykes, and 1.6 for cinder), the anticipated raw coal content of the R.O.M. product was calculated at 25m or 30m intervals, roughly corresponding to the average weekly ad vance rate of a longwall face under difficult cond itions. The graph thus obtained (dotted line in Figure 13) was coupled with that established for the "laboratory yield of clean coal at 1.50 RD" (dashed line on same figure) . For any given position of the face, the predicted "theoretical washery yield of the R.O.M . product" from this longwall was obtained by multiplying the raw coal content (dotted line) by the laboratory yield (dashed line). The result is illustrated by the full line graph in the same Figure 13.

SPECIFIC STUDY OF THE DYKE ZONE The formulation of the above predictions impelled the Mine Planning Department to consider two unpleasant options: (a) Go ahead and extract the whole block according to the initial plans. (Corollary: suffer the consequences in terms of low recoveries affecting the budget targets, very difficult roof conditions and incalculable delays); (b) Leave the middle section (a 227m stretch) unmined . Drive a new face heading, and mine the southernmost portion as a separate longwall block. (Corollary: time lost in moving the equipment, lost production) . The geologist was requested to submit his views on the problem in terms of tonnages and percentages, the final decision being postponed to await the outcome of the argument which was to follow between mining engineers, responsible for the fulfilment of production targets, and coal proces sing engineers , responsible for the preparation of the saleable product. To mining engineers, the first option seemed more attractive, though its only chance of being successful apparently depended on a very high rate of ad vance through the disturbed zone.

The 0.5 to I .Orn vertical displacement along the southern dyke was likely to cause roof and floor control problems impeding the projected advance rate, so that, in the worst case, the crossing of the zone might take much longer than the time needed for moving the equipment to a new face heading. But the engineers were confident in the robustness and the good performance record of the W-SD unit, and the potential loss of some 81 000 tonnes of coal to be left un-mined weighed heavily in favour of the first option. This coal, it was argued, was more than enough to compensate for any loss of time which, probably, would not occur. From the viewpoint of the coal processing plant, low recoveries, high ash and a low C.S. Number, or even the overall tonnage of refuse were only minor inconveniences compared to the nature of the igneous rocks, which were totally altered to clay. It was emphasised that about 4 to 6 tonnes of this in a single shift would mean every single hole and pore being filled with fine clay. A general clean-up would be needed after any such shift, without mentioning the substantial amounts of good quality coal that would be thrown out as refuse. To maintain the washery at its optimum level of efficiency, the transit of even smaller amounts of fine clay should be avoided at all costs. The graphs on Figure 14 indicate the overall tonnages of non-coal material in the R.O.M . product, from the position where the face reaches the dyke / cinder zone to the position where it leaves it behind . It also shows, in terms of "tonnes per metre of advance", the composition of the stone component of the R.O.M . product. For a face advance of 25m per week, for instance, up to 50 to 55 tonnes of fine clay per shift would be present in the R.O.M . product from this panel. It was calculated that this amount was likely to vary between 15 and 30 tonnes per shift for the washery feed , while the face went through a 140 to 150m section. For an additional 55m zone, the amount was likely to remain above the critical level. The only way of maintaining the clay tonnage at tolerable levels, it was concluded, was to achieve a much slower rate of advance. In this case, however, the crossing of the zone would take a minimum of six months (totally unacceptable), without mentioning that a deliberately slow pace under a weak roof would render roof control problems more acute, and thus slow the rate further and further down . On the other hand , washability and ash curves (Figure 15) were worked out by the geologist to illustrate, independently of the effects of clay in the washery feed , the float / sink characteristics of the coal and of the R.O.M. product: (a) Supposing the whole block was to be mined; (b) Supposing the dyke zone was to be left unextracted Theoretical yields for specific washing medium densities, and "ash content against yields" were plotted on two sets of curves corresponding to these two options. A difference of 8% in overall yields for this panel alone meant a potential drop of 3% in budgeted plant yields for the year 1978/79.

103

> <t

AS "H

CS.N.

RE"TREAT

Figure 12

Predicted ash content and crucible swelling number of washed coal at 1.50 R. D. for various positions of the face line.

SU - BLCfK TI EE EX TRA TED

l•foce 202- ~)

;

/I \ \,_

,./

UB -E LOCK TO B EX RACT D ( ,Fa e 2( 2-A)

!===t===l===t===l==ct==t==r' ~ - - ~H1==l==l==F=t==:l==:.l==tr-t't,---:i-t-111il-i=+====!===t==== \' , I

\\...._ /

-l~'~---+--l---h...-nl...--rd.=--+--+--++--l---"l''----+---+-+-+---+ - :j&-L--4L--+--l----!--l--+--+,ca, Lt <m• ,es; SU3-'IECTICN TO E LE ~T ,v

Figure 13

I UNMINED,

21 0

100

!IJ

Predicted variations of the R.0.M. yield at 1.50 R.D., for proposed extraction heights ( = theoretical washery yield).

104

MODIFICATIONS OF INITIAL EXTRACTION PLANS, BUDGETING On the basis of the advice given by those responsible for the coal processing plant, and by the mine geologist, the idea of mining block 202 as a whole was abandoned. Two economically mineable sub-blocks, named L W 202-A (north) and L W 202-B (south) were identified. The washability and ash curves in Figure 15 (i.e. set "b" only) were then combined with those prepared for other panels in the Bulli and Balgownie seams, in working out the annual budget proposals for washery yields. Three additional sets of curves were also prepared for the two halves of block 202-A, and block 202-B to indicate yield and ash targets for quarterly periods (Figure 16). The main purpose of this was to attract the attention of mine planning and washery personnel to likely quality fluctuations in the course of the budget year. According to these predictions, the L W 202 component of the "South Bulli R.O.M. product" would allow a gradual increase of the washing medium density. In other words, the deteriorating impact of the component was likely to be felt very acutely at the beginning, but this influence would gradually be transformed to a beneficial one towards the end.

EXTRACTION MONITORING

THE

200m to end: Spectacular improvement m roof conditions due to a high speed of advance. Cutting height reduced to suggested level.

.--"'~,---··' •••;:..,., _

.1::-f) - ....

'11:!1"-c;,-

,,._/'

tolClll'OOl tOC - f

111JIIIT

t,f11.E\_ & 'i,!Ll.'i

,,.•-•··-·

: • • -cl...,.. ,1_ ,ol,t

;

Tokll s -·•W, n l• - !.cl lDU 101

... .... _...._., 1' 1Pt • -

· · - ~·

.r!

Ji ..

i6'

Dyke zone: Analysis of the anticipated non-coal component of the R .O.M. product

Figure 14

'IELI

,,- / 1~v _, / ~-

o/, 0-

Q · lj{_

,,,"

BLOCK,

No major problems were encountered during extraction of the block, other than some unexpected roof falls. All mining characteristics, i.e. extraction heights and dilution rates, were monitored at regular intervals. The actual performance is compared in Figure 17 to the initial predictions of the mine geologist, in terms of "yields of the R.O.M. product at a given washing medium density". It immediately appears that the predicted overall yield of 78.3 % is practically identical to the actual overall yield calculated after completion of the block, which is 78.0%. However, very important discrepancies occurred between the predicted and actual yields along ome subsections of the two blocks, for the following reasons: BLOCK 202-A 25m to 145m from start: Dilution from roof less than anticipated; 145m to 315m from start: Extensive mining of floor rocks at tail-gate side plus fall at tail-gate between 155 and 280m; 280m to 503m from start: Roof conditions better than expected. Increased yield partly due to reduced height of extraction (2.4m to 2.45m instead of 2.6m as suggested). BLOCK 202-B 65m to 200m from start: High rate of dilution from roof (sandstone closer to top of coal than ex pected). Extensive mining of floor due to height of extraction being kept above 2.7m;

2t'SIXI I

,1 i/ // ~

.........

,- b

,,, ,

...u _ ::,

b->-

, I/

1 V

f L-~

a: -a

,,o-

,. t

-·I

Figure 15

1/.'

l .. I

t i5

Washability and ash curves (a) including dyke area; (b) excluding dyke area.

l60 SG

c....J.• ·- - - - 1 ~ AT

105

GEOLOGICAL WOR K AFTER COMPLETION OF LONGWALL

u.!c..r~

-+----+-----+-"'-1--+-_._---+--< TT'HO ---l-+--4--+---+---+---1-1----

i --+--+-l--++-+--+--+--+----i-1--+--+ ---+--

\ .,_ I +--+-+---1- i --+--+-l--++-+---il~

Observations made during the development of the block, together with the results of face monitoring provided invaluable information for the development of the remainder of S-District. The presence of a dyke system, first intersected in SI and S2 panels and apparently widening in a southwesterly direction,had been rightly considered as a potential threat to the development and extraction of further longwall blocks according to the initial district lay-out. The preparation of alternative lay-outs had been advised at an early stage. Such major changes in mine planning, unfortunately, cannot be achieved without paying a more or less high price in lost time and lost production. Thus, the next block (i.e. L W 203, west of S2 panel) was planned, from the start, as a double unit comparable to LW 202. However, as pilot headings indicated that the widening dyke zone also included a teschenite pipe and a massive, unaltered sill, a totally new lay-out was finally adopted for longwalls 203 and 204, circumscribing the dyke / sill zone without ever crossing it. Monitoring of face conditions also indicated that a certain additional percentage of dilution from roof rocks should be included, as a safety margin in yield calculations, to the tailgate side for future longwall blocks.

=c·

/,_..._-+---+-v--+._,,.,... t-

-+--t--+---t+-l-\-+t--+--+--++-+---tf--+-- -+-+-7"f--b,....ql0l·• J -

X)L,, , _

\ \

10r•-,··

... ···· a:,:..

... \ ..... .

___ _

.. ,,.

-+-l--+--+--l---4---1---<e-+---+- 1-+---+--+-

Washability and ash curves of various subsections scheduled for extraction

Figure 16

' I

t -

I I

- · - - PRE Ol( t'EO YlfLO

-!'CT\JA

I _I

f Froin Tobie 7 J

(CG\a:' la~I 1YIELD

-·

\._ .,,..\

.1---- s--

, :.;=a .•

~ l!l -

I\-,--

-.-

i:""' '•

11\

._..,--

_.,.,.-

....

,.__--

I-Bi

,/"..

I~ \ ~

:::,

·y

----

~ ' ••,

! ;

; I

.,._

•Ill

Figure 17

ACE 202-

200

lino

,....... ,do

. .

V \\

-- l'li.

w a:

.... -- ,... AC;f;_

)

500

Ji,,

JOo

02-A (,,,..trdl

,&,

,do

... , _,-.r,_

w -'

~ -

n 1

300

srk

Comparison between predicted and actual yields of the R.O.M. product at 1.50 R.D.

106

CONCLUSIONS The problems faced in this instance point to the vulnerability of detailed early planning in a virgin coal domain earmarked for extraction by longwall mining. Although accurate predictions can be formulated for coal quality from borecore data, errors are very likely to occur in assessing the trends on a directional or on a time basis. Borehole data is often inadequate for predicting roof deterioration due to lateral variations of roof lithology, jointing and stone-rolls, but predictions based on accumulated experience are reliable on an overall basis. Most dykes, in the area considered, do not reach the surface, and being too thin and , in most cases, totally altered to clay, cannot be located at a distance allowing an early modification of the mine lay-out , by surface or underground geophysical techniques currently in use . Compared to bord and pillar methods, the longwall system is too rigid for an immediate change in extraction plans, without considerable delays and associated production losses. However, very detailed geological work in development panels is of paramount importance in indicating ways of alleviating the potential impact of an unforeseen feature on the overall R.O.M. product and on washed coal characteristics. Short-term predictions of coal quality and mining characteristics proved to be most reliable, whenever suggested shearing heights were adopted and respected.

I07

Coal measure geology and the planning and operation of open strip coal mines Graeme Boyd CSR-Energy Division, Box 650 G.P.O. BRISBANE, QLD 4001

ABSTRACT It might be claimed that ideal open strip coal mmmg conditions are those that allow optimal use of excavation equipment in handling the minimum amount of overburden for the maximum amount of coal uncovered . However, many factors frustrate this ideal, significant among which is the inherent make-up of coal measure sediments. Specific but pervasive properties of these sediments have been analysed in an attempt to overcome such operational problems as mass instability of pitwalls, optimisation of blast design , assessment of ground excavation characteristics for deep strip mining as well as the assessment of ground conditions to provide baseline criteria for use in feasibility studies of new prospects. Relevant properties of the coal measure sediments include clay mineralogy, the nature and extent of secondary cementation, the attitude and orientation of bedding planes, the thickness of lithological layering and the geotechnical characteristics of the bedding surfaces. These properties in turn are related to early diagenesis of the sediment, post-depositional deformation of the sediment stack, the extent of weathering in the Tertiary, and seam displacements as a result of faults developed at various stages in the lithogenesis of the basin sediments. These relationships are described collectively in an account of the approach, methodology and techniques used in the prediction of ground conditions, and hence, operating conditions, ahead of mining.

INTRODUCTION This paper presents, by way of a case history, the techniques employed or currently being developed to measure inherent and induced properties of coal measure sediments likely to affect the design and operation of large open strip coal mines. The reason for such development was born with the need to overcome a severe pit-slope instability problem experienced at Central Queensland Coal Associates' Goonyella mine, northwest of Mackay, operated by Utah Development Company. Here the overburden is first blasted , then stripped down to coal by 45m 3 capacity walking electric draglines. The coal is then blasted and mined by an 11 m3 electric shovel or a 14m3 front end loader before trucking to the preparation plant. In this type of open strip operation, several geological factors may affect the design of blast patterns to improve fragmentation , or the conditions of highwall (i.e. hanging wall) stability. The stability of the spoil dumps is determined by the properties of the overburden materials and the manner in which the dumps are constructed, as The well as by groundwater conditions. overburden stratigraphy, the degree and extent of weathering, and properties of rock mass discontinuities are fundamental to highwall and lowwall (box-cut) failure mechanisms. In contrast the extent of soft, easily dug overburden is a key factor in determining ultimate open strip mine depths, particularly where the application of bucketwheel excavators is envisaged. Recognition of these geological factors which affect orderly open strip mining operations at Goonyella resulted from detailed geological studies

of the highwall, and of the performance of associated natural materials during the repetitive cycle of blasting, stripping and spoil pile This knowledge is now being construction . extended into prestrip areas ahead of the operation, with the need to develop new detection and measuring techniques for acquisition of adequate data to be used in all phases of mine planning.

OPEN STRIP OPERA TI ONS IN COAL MEASURE SEDIMENTS A number of inherent and induced properties of coal measure sediments contribute to adverse mining conditions, at times on a scale that may severely influence the viability of the stripping and mining operation. Problems that can be directly ascribed to properties of the exposed and excavated coal measure sediments are a follows :

Operational problems related to ground conditions Pitwa/1 instability Comprehensive investigations of the high incidence of pitwall failure at Goonyella concluded that pitwall instability resulted from structurally controlled deformations along (essentially) planar failure surfaces. In most instances, the failure surfaces were generated as a direct result of the manner in which the slopes were constructed, and from the adverse affect of blasting on key structures in the overburden. (a) Spoil dumps. In spoil dumps, the structural zone of

108 weakness was found (Boyd and others, I 978) to coincide with : (i) weak, planar surfaces in the surfaces in sub-floor strata, or (ii) a weakened (softened) layer of spoil material in the base of the pile. The incidence of failure due to weakness of the sub-floor strata dim inis hed with successive strips, despite the fact that slicken sided and / or polished bedding surfaces in the shaly seam floor continued down-dip. The majority of fai1ures· however, occurred due to softened and weakened spoil material lying in the base of the dumps themselves. Figure l(a) demonstrates the reversal of material strength that occurs towards the base of the spoil dump. Figure l(b) describes the varying stability (measured by the factor of safety, Fs) as a function of the area (length) of the base plane made up of material of variable shear strength (i.e. variable <j values). The base case for comparison (cohesion = 10 kPa, ~ =6°; Fs = I) is representative of (total) shear strength parameters for softened and degraded spoil lying !n the base of dumps. (b) Highwalls. The most significant structural factors that caused instability in mine highwalls were found to be: (i) extensive surfaces, corresponding to sheet bedding, in laminated to thinly bedded strata; (ii) strength anisotropy, due to the welldeveloped bedding and the weakness of bedding in certain strata; (iii) the occurrence of subvertical joints, usually forming orthogonal joint sets; ( iv) blast-induced tension cracks or extension / dilation of joints. A "two wedge failure " mechanism was interpreted from subsurface instrumentation, with the "passive" or front wedge sliding over a generally planar surface, while the "active" or rear wedge deformed in a complex manner, probably assisted by separation of the rock mass into upright slabs and columns. For kinematics to be satisfied in gravity-induced slumping, the planar surfaces usually possess a reverse (non-conformable) dip towards the pit. In such a case, although the dip of the coal seam may, for example, be to the east, the bedding in the strata forming the overburden of that seam would exhibit a dip component towards the west. The degree of instability in these circumstances is directly related to the amount of dip reversal (see Figure 2). • The origin of structures causing such dip reversals was investigated in some detail by the author (Boyd , 1980). These structures were found to be associated with postdepositional folding, usually found in association with the development of seam splits. (c)

Lowwalls. . Lowwall instability in box cuts also involves a two-wedge failure mechanism, reflecting the

200

Shear Strength (T) ( kPa ) 400 600 800

o ~--~--~--~--~~- ~1000

••

c. .,

• Floor

•

25 '----- - - -- - -- - - - -- - - '

Figure J(a)

Change in spoil strength with depth

20 Strength of toe reg ion

t 18

...,~/",.'1 , iQ / 1/Q / / ~ 7..p_,,%1•

W idth

.' '

36°

30'

,;,

-,•

• W

,,,✓

.,,,..,,, ,,, C • 10 t • s• ] .....:;;~ ~,,,~f~:::~=-- --0---- ..£...:.'..'Llc..:.!~J

08 ~ 0

- - . - - - - r -- - - . -- - , - - ---.----1 10 20 30 40 50 W odth of Strengthened Toe Reg ion ( Metres )

Figure J(b)

Spoil base conditions for stability.

strong structural controls of weak surface(s) oriented parallel to the coal. Such a surface may be above, within or below the seam exposed in the lowwall. An added factor of some significance in this situation , however, is the geometry and position of the boxcut spoil emplacement relative to the lowwall crest. Thus, apart from the inherent instability of the weathered overburden at the lowwall, the risk of failure is increased by loading the lowwall unduly with boxcut spoil, as indicated in Figure 3.

109 150

100

"'"' ~

cii

i 50

.r:::.

C/l 40

·J o 20 10 Q 0<-~~~-~~~-~~~-

Bedding Dip ( Degrees)

Effect of dip on highwa/1 stability

Figure 2

Ruulr lo, f1 • ed • lo w w all height • spoll dump tuught

c_____,""--

-)Fsm,rl1ma

• co•lsec1md1p

\umque fs m1mmc1 tor each bo l Ct.JI geometry W-(m )

Figure 3

Boxcut geometry versus stability

Groundwater It is usual in coal-bearing sequences that the coal seams themselves represent the major aquifers, with the permeability of over-lying strata being a function of the tightness and the frequency of the joint patterns that intersect to the seam. Similar "structural" permeability in the floor strata may result in seepage of water into the pit, or worse, into the base of the spoil dumps. Groundwater thus becomes a problem where: (i) pit access and trafficability is impaired; (ii) contact with water-sensitive spoil materials is difficult to control; (iii) significant hydrostatic forces develop behind the highwall or in the pit floor. Runoff from monsoonal rains in the immediate pit catchment commonly results in the buildup of large quantities of pitwater, with depths of up to 10 metres, and this clearly has major implications to pit operations and pitwall stability.

Overburden blasting Good fragmentation of the overburden leads to optimum productivity and cycle times from the draglines through high bucket fill factors and reduced digging (see Table 1). For bucketwheel excavators employed either in prestripping operations or for removal of prime overburden, fragmentation is critical not only for system productivity but also for its total capacity. Oversized fragments not discarded at belt transfer chutes cause serious damage to conveyor belts and idlers. Fragmentation is closely dependent on compositional and textural variations in the sediments where such properties influence material strength anistropy. The structure of the rock mass, in terms of the spacing and properties of discontinuities, can now be modelled by numerical methods in computer simulations of specific blast designs (Harries, 1973). Poor blasting practice has been identified as a major cause of pitwall failures. Blasting induces the following changes in rock properties: (i) the generation and deformation of firsttime bedding plane fractures. These are commonly developed at boundaries of modulus contrast (i .e. boundaries between hard and soft materials), and where exposed bedding surfaces are water-sensitive and hence soften on groundwater ingress in the highwall zone; (ii) the development of upright surfaces of separation, which may coincide with original or extended joints. These represent "release of load" structures (Hagan and others, 1978), developed by removal of compression from the rock mass in the zone of back break. The gas generated by the explosion vents rapidly and thus peak explosion pressures are reduced. (iii) the creation of a block-jointed rock mass in the zone of back break (i.e. in the new highwall), in which the degree of confinement of the blocks is significantly reduced . Since the bulk strength of the rock mass is a function of the degree of confinement, the amount of fragmentation in the zone of back break has a significant influence on highwall (or lowwall) stability, particularly under the hypothesised failure mechanisms associated with active wedge development (Cox, 1979). Pre-stripping Dipping seams relative to the land surface result in incremental increases in the volume of prime overburden to be excavated per strip. More particularly for dragline operations, the maximum depth for "practical" digging is about 40 metres, and hence, second-generation stripping equipment is required when either excessive pre-stripping volumes occur, or when pit geometry of limiting design is exceeded . Ground conditions affect the efficiency (productivity) of pre-stripping, depending on the tee:hnique adopted. The use of large bucketwheel excavators (as intended at Goonyella) demands adequate provision for instability at the digging face, of the ground supporting face or end conveyors, and of the spoil dumps supporting

110 I. Digging hours/ year/ dragline - 1977 average = 6070 2. Swings per digging hour - 1977 average = 53.7 3. Average bucket load = 37.4m1 4. Average swing cycle time = 67.0 seconds 5. Assume cycle improved = 5 seconds 6. Additional 0B moved by 12 draglines per year. = 25,500 X 37.4 X /2m1 = l l ,500,000m1 = I dragline (@ 128 M each) Table I

Optimising dragline performance

spoil conveyors, the spreader unit and the spoil from pre-stripping placed by the spreader. Prestripping by shovel and truck usually requires a terraced working face, the overall slope of which must be optimised for economic reasons, whilst local terrace batters must be designed with regard to safety and minimal interruption to haulage. otwithstanding the pre-stripping technique finally adopted , knowledge of ground conditions and material properties remain common to all. This is particularly so where pitwall instability, fragmentation problems and dump slope stability are important factors in optimising the shape of prestrip areas, yet still allowing efficient mucking and hauling.

Causal factors - geological associations Properties of coal measure sediments identified as contributing to difficult operating fall into three groups: (i) those related to the chemical stability and mechanical strength of the sediment, which in turn depend on its composition and texture; (ii) those related to the usually well developed, persisent and generally planar layering of the fine grained sediments , particularly those situated adjacent to the coal seam; (iii) rock mass discontinuities, including joints and faults, as well as monoclines and roll structures. The following discussion draws heavily on experience with Utah's operations in the Bowen Basin, and as such provides a generalised description of the interaction of coal measure sediment properties and open strip mining. Chemical stability and mechanical strength The result of pitslope stability research at Goonyella clearly demonstrated that the fundamental cause of spoil dump instability was either: (i) very weak bands lying in sub-floor strata or (ii) degradation by crushing and slaking of the spoil located at the base of the dumps. Weak bands in the floor generally occur only in shallow strips, as an expression of near-surface weathering effects. Since successive spoil dumps continued to fail at progressively greater mine depths, a failure mode involving large shear displacements through the material lying at the base of the dump was therefore indicated .

Despite the tendency for large blocks (up to 1m) to accumulate at the base of the dumps during spoiling, analysis of failure mechanisms clearly suggest that failure occurred through the presence of moist, fine grained soils. Such materials were, in fact, recovered by coring the base of the dumps, consisting of sandy, gravelly clays to clayey gravels in which the gravel fragments commonly exhibited features indicative of partial breakdown. Thus, mechanisms involving the in-situ degradation of rock-fill to soil (in the engineering sense) under the weight of the dump, and usually in the presence of water, is believed to be the progenitor of spoil dump failures. Table 2 summarises compositional and textural properties of the major lithofacies usually found in coal measure sequences. The composition reflects matrix material found in coarse grained sediments, or constituents making up fine grained strata. From a compositional viewpoint, the presence of significant amounts of clays with moderate to high swelling potential on moisture absorption lowers the rock's resistance to physical breakdown . In particular, the concentration of montmorillonite in weathered sandstones has a significant influence on dump stability. This influence was shown by Godfrey ( I978) and later confirmed statistically by Lazlett ( I 979) as the reason why locations of recurring spoil dump failures along the mine lay geographically opposite sections with (upper) highwall strata composed of medium to highly weathered sandstone. These sandstones exhibited a greenish discolouration due to weathering and the associated formation of montmorillonite in the material. Whereas the resistance of sediments to degradation may be related to sediment composition, the · extent of this breakdown involves textural characteristics and / or the absence of diagenetic cements. This means that the engineering behaviour of the materials is controlled essentially by the properties of the sediment matrix. As an example, granular sediments found in the German Creek Formation at Peak Downs, and overlying the Baralaba Coal Measures at Moura and Theodore, possess grainsupported textures in which exsolution at grain boundaries has generated a well-cemented granular skeleton. A similar effect occurs at Goonyella in siltstone and some sandstone beds, where carbonate diagenesis has produced wellcemented beds that are relatively resistant to slaking, despite their composition (Table 2) . Sediment layering Sediment layering here includes all types of bedding surfaces (horizontal bedding, crossbedding, trough bedding etc.) as well as surfaces of local erosion or non-deposition. Implications for stripping and mining are related to: (i) the attitude and orientation of the surfaces relative to the strike of the pit; (ii) the inherent or induced fissility of the layering; (iii) the engineering properties of the fissile surfaces; (iv) the sharpness of the layering, particularly where the rock properties vary across such

Ill UNIT I

LACUST1tI,E MUDS

CHA~:,OEL

sA,osro, E

OVERBA'K DEPOSITS

SH[ET

SA',DSTO'E ~Ius1O,E

, DRAPE DEPOSITS

•

COALY BEDS

STRUCT RE

CLAY CO M POSITION

TEXTU RE INTERLAMl'lriATED C20m• SPACl'OI ro THl:,,l, l"'ITER,BEOOEO (60-200mm SPACltloG) SILTSTO ' E· WITH Ml'OR SA!\DST0'[ A'DCOALY Cl AYST0"1E BEDS, DIPS CO'i'FORMABLY WITH COAL SEAM AT J.-1 DEGREES

I Llll E-MO:-. TMORI LLO' I TE

FINE TO MEDIUM CiRAl'ED TROUG H A'D CROSSBEDOEO SA/1,DSTO, E BEDS · THICK la.600fflmtTO MASSIVE (> 2000mm). DIPS ARE RA,DOM , CROSS BEDS DIP UP TO 25• LOCALLY

MO' rMORILLO:O.ITE ILLITE IC.AOLINITE QUARTZ F-ELOSPAR DOLOM ITE CHLOR ITE

INTERLAMl'ATED TO MEDIUM ll'loTERBEDD[D (200• 600mm) CLAYSTONE SILTSTONE WIT II Ml'OR SA DSTO,, t (CLAYEY) BEDS, D I P WITH COALSJ"AM OR AS MODIHED BY DOMAL WARPS W II ERE DIP A',GI [S VARY B(TWEEN 5 A"ID IS DEGREES

H"IE GRAINED (LAVEY SA osro,e A"ID s11.1sro,r lHICIC.LY BEDDED WITH LIGHT A',O DARK GREY C'LA) RICH PARTl'GS LAMl"IATIO,S. BEDS DIP WITH SEAM OR AS MODIFIED BY OOMAL WARPS. DIP A,GLCS VAllY BETWEE' 5 A~O U DEGREES

QUARTZ KAOLINITE ILLITE CHLORITE

.tO . 1, q

"7-. "12 c; ~

,. , . ', "..

20 S

LOCAL CHA'-'EL Fllll ' G DEPOSITS Of LIMITED STRIKE LE,GTH J'I. PIT OR AS POl'T BAR DEPOSlfS ASSOCIATED "-ITH CHA''fl AC'C'RETIO,

12 q 1 q • q

TRACE

ILLffE . MO1'TMOR I LLO/lilTE ILLITE QUARTZ KAOLIMTE MONTMORII LON ITE FELDSPAR CHLORITE

SS- 71 'l I 2- n q.

ILUT[ • MO',TMORILLOMTE MONTMORILL0"1E ILUTE K.AOLl'ITE QUARTZ fELDSPAR CHLORITE

41 - 7S q 1- 29 q 3 · I6 'i 5 . 5f t; I • 2'> Ci I C, TRACE

2 . n q

J - 17 '1 0 • 7 ~

, _, 'i

RH,TII M ICALLY BEDDED t\ ARVED APPEARA ',('[J LATERALU PERSISTE,r . OFrr, TRU ' C'ATEO Pl'CHED Ol'T. SUBJECT TO LATERAL MC'IES C'HA,or WITII (4) BELOW

1-RACE

THllloLY LAMl'ATED f<:6a:im). GREY TO DARK BRO'A•"I CLA YSTO E. BEDS DIP WITH UhDERLYllloG STRUCTURES. DIP A'GLES ARE BET\l,£E"I 3 A'D 6 DEGREES

ILLITE • MO,IMORILLO,JTE !LUTE KAOLl,ITE QUARTZ

SEAM SPLITS WITH ASSOClATEO OVER A"'i'O u,oER CLAYS, DIP WITH SEAM OR MODIFIED BY OOMAL WARPS. DIP AI\GLES ARE BETWEEN 5 A'O U DEGREES

Cl.A'! COMPOSITJO:,. AS fOR DRAPE DEPOSllS

Table 2

SHE£T BEDDED. co, FORMABLE \lr, 1TH COAL. \\ 1TH POLISHED A'D SLICKP,.SIDED BEDDl'G A' D OH BEDDl'G SHEAR PLA,ES

"' "' '""'

I4 t;

RH)l HMICALLV BA,DED. GR.-.OED C'OARsr l 'P"ARO LATERALLY PERSISTE , T. Sl BJECT TO IAHRAL FACIES CHA,GE WITH 431 ABO\ E

BAS£ ASSl'MES TOPOGRAPH'I OF SLRfACC TO l ,oERUl,G MATERIAL LAMl'AE ARE SHEITlllE. LATERALL'I PERSISTE , r

LATERALU PERSISTE-.T. DELl :r-r. EAlE~ OOMAL \l,ARP STRl'CTlRES. OCCASIO,ALLY FALLTEO

Description of dominant stratigraphic units

surfaces; (v) the effects of the layering on fragmentation . As stated previously, pitwall instability in coal measure sediments occurs largely due to structurally controlled deformations along (essentially) planar failure surfaces. For lowwall failures , bedding surfaces conformable and adjacent to coal coincide with the (basal) planes of failure . By contrast, highwall instability usually occurs along non-conformable bedding surfaces with reversals of dip, thought to be related to postdepositional warping of the sediment stack (Boyd , 1980). Other than their attitude, the ease with which the bedding surfaces are separated or sheared is Fine grained critical to pitwall stability. sediments, in which the layering may form thin laminae ( < 6mm separation) to thin beds ( < 200mm separation), usually possess a strength anistropy (ratio of strength normal to and parallel to the layering) of between I and 6. In sandstone, this ratio varies between I and 4, but is most commonly between I and 2. Introduction of diagenetic cements or change to a grain supported fabric reduces this strength anisotropy for all materials. Given that fine grained sediments usually have a moderate strength (0.3 to 1.0 MPa on an Is 50 basis), the relative strength along sedimentary layering is low. This promotes an induced fissility in the pitwall as a result of either: (i) overburden blasting; (ii) dissipation of elastic stresses on excavation; (iii) action of picks during digging. Once separated, bedding surfaces expose clay-rich materials to seepage, particularly where blastinduced fractures are propagated into the rock

mass at levels below the standing water table . Any moisture absorption may result in partial degradation of the fracture surface to soft clay, with an attendant reduction in shear strength of the fracture zone. Such a mechanism was clearly evident in one very large highwall failure at Goonyella , involving over I million cubic metres of material. Rock mass properties controlling fragmentation include the spacing and persistence of pre-existing discontinuities, and the orientation and spacing of the fabric elements in the sediments that control strength anisotropy in the rock mass. Preliminary results of blasting research at Utah's Blackwater mine indicate that, from standard square-in-line blast designs using ANFO 96, the resultant muckpile included blocks with aspect ratios (maximum chord length to thickness ratio) varying from I to 9 (see Figure 4). However, the more sharply defined the bedding, as in thinly interbedded sequences of two rock types, the more pronounced the slabbing developed (i.e. smaller block thickness at a constant aspect ratio). Fragmentation is difficult for dragline operations in massive sandstones carrying persistent, widely spaced upright joints, and in which weak layering is associated with an anisotropy index of less than 2. Excavation of the muckpile by shovel or front end loader, given the results in Figure 4, becomes increasingly difficult. For bucketwheel excavators, where 100% passing a 500mm lump size is a typical specification, the necessity of detailed knowledge of material properties, including bedding separation, is vital for the design of blasts to yield manageable fragment sizes. Discontinuities The pre-existing discontinuities found in the rock mass and their influence on the behaviour of pitwalls are as follows:

112 M «h•n•~t ,.1,0

Cl Ill C Cl ·- C

undstonr

-C · -

-ct Q)

\ \

.: V ••

S•ndstone ...,,,,. ,,,.,_ wHH.-

SLA B ASPE CT RATIO

10 15

~=,:~~,=.: c::~o:::'ff

ID~

.c. +a. 10 Q)

Figure 4

Fragmentation results for interbedded strata

(i) bedding plane shears or partings. These are most prevalent in the sequence as surfaces of weakness in fine grained, laminated to well bedded sediments. Their prevalence in such strata adjacent to coal seams, as shown in Figure 5, suggests a post-depositional origin; (ii) low angled curviplanar joints, often carrying polished or slickensided surfaces. These commonly connect and thus extend bedding shears or parting further through the rock mass (Figure 5). They are thought to be of post-depositional origin, as for the sheared bedding partings; (iii) faults, either normal or overthrust. Overthrust faults influence wall stability either by forming a surface on which gravityinduced sliding is initiated or by folding adjace nt sediments to form surfaces of potential sliding. Upright, normal or reverse faults , by comparison, cause minimal impact on wall stability. However, difficult operating conditions are created where 2m to !Orn displacements along fault sets form local graben and horst structures, resulting in difficult recovery of the coal and disruption to all phases of operational scheduling in the area; (iv) upright joint sets. These are generally normal to bedding and persistent in massive strata , but offset and limited to one or two bed thicknesses in finer grained or interbedded strata. The influence of these features on failure mechanisms is not well understood, but they do provide planes of separation during the formation of subvertical columns associated with shear displacements along surfaces of sliding. They also have indirect control on stability where open joints impound surface runoff behind pitwalls, causing both seepage and softening along blast-damaged bedding surfaces, and hydrostatic forces on blocks to the pitward side of the joint, adding in turn to the forces promoting failure .

Targets for geological mapping The targets for geological mapping described below are related to the major problems of pitwall performance, blast design and excavation

....

20 Figure 5

Structures associated with coal seams.

characteristics likely to be experienced at some stage in optimising operations at any one mine. Other difficult conditions may also arise, such as those caused by groundwater inflow or significant structural disturbance to the seam, but based on general experience in Australia to date, these are likely to be site specific. A programme that sets out to define the material or substance properties, the characteristics of the rock mass, the groundwater conditions and the features that may deform under the expected induced stress should enable reliable prediction of stripping and mini ng conditions and hence allow the development of the most appropriate pit design. Material targets If strata are present that have a potential for rapid breakdown on moisture absorption or for excessive mechanical crushing, detailed mapping is required to provide a basis for scheduling t he stripping operation and allow stable spoil dump construction. On the other hand , if large amounts of mechanically resistant, strong material are present, in which the inherent planes of weakness are poorly developed and natural discontinuities are widely spaced, particular blast design or equipment selection may be necessary. Sediment compos1t1on, texture and related controls, as observed directly or interpreted from index (e.g. strength) tests provide a comprehensive basis for delineating problem areas. The extent of adverse sediment masses and their locations in the sequence need to be defined and hence mapping of the following geological features is warranted: (i) base of the superficial soil cover; (ii) base of the weathered zone in coal measure strata; (iii) stratigraphic position and extent of the major lithofacies, as described in Table 2. Defining the state of weathering associated with poor slaking or crushing resistance depends on local factors involving the original sediment

113 composition and the weathering environment. As a general rule, the base of the moderately weathered zone may be mapped i.e., the zone in which significant alteration and discolouration around joints pervades into the rock mass. More subtle evidence of weathering may have a less than subtle effect, particularly in sandstone strata. Hence, in at least one location, the composition of the rock in each weathering state should be investigated by x-ray diffraction analysis. Swelling clays, particularly montmorillonite or mixed-layer illite / montmorillonite minerals deserve special attention in such studies. In fresh strata, the clues to rock behaviour are not so clearly evident. In coal measure sequences, fluvial, upper deltaic plain or lower delta depositional facies, levee and flood plain strata comprise sediments reworked by numerous shiftings of meandering channels. Thus, the sediment compositions are well distributed, both areally and in size fractions from fine to coarse. However, where adverse compositions and / or textures are recognised, some stratigraphic control will probably exist. The extent of these problem strata can be determined by lithofacies mapping of highwalls or by detailed correlation between boreholes. For example, a lithofacies mapping programme that defines the location and extent of massive channel sandstone deposits will assist in identifying areas in which blasting and / or fragmentation problems are likely to be met (Mallett and Flood, 1979). Extra benefits of compiling this data relate to assessing the viability of alternative stripping methods (i.e. dragline, bucketwheel or shovel / truck) over the planned life of mine, both fot prime overburden removal and for prestripping requirements. Again, the base of the various weathered zones must be determined to suit the digging characteristics of each method. Structural targets The potential for pitwall instability can be anticipated by focussing on the following structural targets: (i} formation dips in boxcut areas; (ii) post-depositional folds in the overburden; (iii) faults. Due to the location of boxcuts adjacent to either the coal subcrop or the limit of coal oxidation, it is not uncommon for formation dips to increase or for the seam to become contorted, and the overburden to be locally folded and / or faulted, in the strata forming the lowwall. The situation is exacerbated where the coal subcrop also represents the depositional margin of the seam, and where steep monoclinal folding and / or post-depositional drag faulting might be expected. Due to the importance of establishing stable boxcut conditions to maximise pitwall stability in subsequent strips, it is prudent to bias field investigations towards boxcut areas. The location of trial pits in obvious boxcut areas, in particular, can yield valuable information for the development of stable mine designs. Any structural feature which renders the orientation of the strata different to that of the coal seam is a target for delineation. In open strip

coal mines, the structures of this type comprise post-depositional folds in the interseam sediments, in which dip reversals of beds in a high wall relative to the coal, may occur over a strike length of up to 3km. Jn existing operations, such structures are clearly visible as folds, with characteristic asymptotic flaring of beds on to a thin fine grained stratum immediately overlying the coal. Since the shape of such features may vary from domes to ridges, it is difficult to predict their extent beyond the highwall without exploratory drilling. Detailed measurements of bedding attitudes at the one level across such exposed structures will generally indicate whether successive strips are moving towards or away from the position of maximum amplitude of the fold (i .e. towards or away from the point about which dip orientation changes). Where drilling results provide the only data available, indicators of such fold structures can be gained from the following features : (i) variations in the thickness of particular partings over short distances; (ii) locations of any coal seam coalescence; (iii) presence of overthrust faulting in or near the seam; (iv) non-conformable attitudes of fine grained strata relative to the coal seam, as interpreted from down-hole orientation devices. The line of seam coalescence generally defines the limit of non-conformable dips due to fold development, and as such provides information as important to the geotechnologist as it is to the mine geologist. About the position of this line, usually just inside the structure, low angled overthrust faults to high angled reverse normal faults can be anticipated , with curvilinear trend of approximately the same trend as the line of seam coalescence. Faults may also be found in the following depositional settings: (i) at the depositional margin of a seam; (ii) beneath the flanks of massive sedimentary bodies such as channel sandstone or conglomerate deposits. The style of faulting may include normal faults, overthrusts, and bedding plane shears in laminated material.

DELINEATING THE "PROBLEM" GEOLOGY The techniques applied to the measurement of specifo; features and to the interpretation of the acquired data from Utah's open strip coal operations have gradually evolved in response to the following difficulties : (i} very large strike lengths (over 50km) of pitwall must be mapped; (ii) approximately 100km2 of ground area, in which stripping and mining conditions must be predicted lies ahead of current operations; (iii) approximately 150 to 200km 2 of ground area is yet to be developed within the current mine lease boundaries.

114 Field procedures and techniques were thus required to give rapid yet accurate data for comprehensive evaluation of large areas by a relatively small group of geologists and geological assistants.

Field techniques The geology exposed m the highwall is an invaluable data source in existing operations. However, as predictions are required at greater distances from such exposures, sub-surface exploration must be increasingly relied upon. Common to both situations is the recording undertaken at Goonyella of objective data by photography, either of the high wall, or of the walls of boreholes. A scheme of rapid mapping of the overburden stratigraphy to delineate important compositional and structural variations was developed that relied upon oblique aerial photography of the highwalls. The general arrangement and relevant details of this operation are shown in Figure 6. In addition to significant savings in direct costs and manpower compared to conventional ground mapping, the following benefits were obtained: (i) a permanent, high quality record of the pitwall geology; (ii) mapping of individual beds down to 25mm thick; (iii) spacing and semi-quantitative measurement of important structural features; (iv) rapid, accurate measurement of structural domains. Repeating the operation for successive strips allows a three-dimensional reconstruction of the geology, particularly the distribution of the dominant lithofacies or rock mass units associated with bad ground conditions. The greater the clarity of this three dimensional model, the further beyond the highwall can stripping and mining conditions be predicted. Down-hole photography was employed beyond the highwall to provide data on the composition and structural properties of the sequence . At the time of writing this paper, the down-hole camera developed principally by the CSI RO to specifications provided by Utah, was undergoing evaluation trials with some success. The chief requirements for the device were to: (i) provide high quality reproduction of a 360 degree view of the borehole wall in which strata and structure type could be clearly identified ; (ii) give measurements of structures, principally bedding, to within I O of attitude and 2° of strike. The sensitivity of highwall stability to dip amount and direction, shown in Figure 2, indicates the need to measure orientation and attitude of structures to the specified accuracy. Surface support for the camera was designed to allow operation by a single, trained technician, thus freeing geological staff for detailed analysis of the information obtained. The degradation for overburden strata was evaluated in the field by a simple adaption of the Emerson crumb test (Emerson, 1967). With experience up to seven

categories of material reaction on breakdown can be assigned to the strata, varying from rapid and complete disaggregation to the fundamental particles, to no observed breakdown nor surface softening, i.e. no moisture absorption. The potential for extracting useful mmmg engineering data from down-hole geophysical logs appears promising at Goonyella at the time of writing this paper. In particular, the neutron log, in discriminating porosity changes, can also indicate variations in the elastic modulus of the material and hence areas where soft or hard digging (e.g. for a bucketwheel excavator) or difficult blasting conditions are located.

Data analysis Down-hole photography As part of the CSIRO system, borehole photographs can be taken at pre-determined intervals, depending on the wall-coverage sought. After processing (a possible onsite function), the film is wound manually through a projector on to a white board, allowing details of the strata lithology etc. to be encoded on to a coding sheet, taking into account the depth of each observation as recorded in the film. Structural measurements are obtained by running the film over a graticule placed in the projector. The graticule is constructed from the projections of planes of known dip viewed through the camera lens, cancelling out any slight aberrations in camera optics. Measurements can be recorded directly on to the coding sheets, along with the lithological information. Highwa/1 photography Oblique aerial photography taken by mine staff is processed locally at Mackay. The film is first rectified to remove distortions of scale laterally and vertically across the highwall face . It is then enlarged to the desired scale, and colour contact prints are prepared. The prints are cut into strips containing the highwall, and, with a 60% overlay, they can be viewed stereoscopically. With a strip photo up to 600mm in length, stereoscopic viewing by means of a Wild ST 10 stereoscope allows complete coverage of individual photopairs. This twin bed viewer also allows coverage of large format photographs through a viewing piece suspended on a travelling arm. Approximately 200 metres of wall length can be mapped in this way from each pair of photographs. Field tests - general Results of any field testing such as slaking characteristics, point load strength etc. can be entered on to an appropriate coding format. Geophysical logs can also be digitized, stored on a nine track tape and retained as a separate data file for interactive plotting when required.

DATA PROCESSING & PRESENTATION Successful application of a solution to any problem is realised only when the mine operations or design staff understand, accept and implement the various recommendations in accordance with

115 Requirement may extend to at least 52km of h1ghwall photography/ mapping per annum 2 Fairchild RC - 2 arc photo camera using premium grade fine grained colour film. mounted in CESSNA 150 with side door removed 3 Photography time ~ 50 minutes for 11 km 1nvolv1ng 10 separate flight paths 4 Film capacity allows 25km of photography for 60% photo overlap for stereo viewing 5 300m of highwall per photo 6 Preliminary analysis of recently completed first trials indicates that compared to terrestrial photographic techniques: • much less scale distortion (less than 7%) • 40% decrease ,n time to complete • 26% reduction in data acquisition cost 7. Sequence of execution · 7·1 M ine surveyors provide wall angles and mine intervals of similar wall angle 7 2 Phot ographic consultants calculate best flight paths 7 3 Flight path indicators layed out by mine surveyors 7 4 Photo runs made in early A M . while wall in shadow 7 5 Contact prints made. rectified if necessary and enlarged to desired scale as stereo -strip photos

FILM PLANE

Figure 6

De1ails of oblique highwa/1 pho1ography.

the specifications given . Accomplishing this result may tax the geologist's skills and mental agility as much as was required to solve the problem in the first instance. This emphasises the importance of clear presentation of data at all times. Not only should the data aquired therefore be relevent to the problem, but its processing and presentation should promote a clear, working understanding of the problem itself and allow the practical aspects of its solution to be realised by all concerned. Where possible, graphical presentation is attempted, which, with recent developments in computer plotting capabilities, allows not only rapid processing of large amounts of data but also gives high quality drawings ready for immediate reporting. However, this apparent convenience also presents pitfalls, particularly where the sophistication of data processing and refining available from computer technology is not matched by the geologists' understanding of the ground being analysed. Geological analysis, by manual drawing of plans and sections if necessary, must be provided to generate dummy or phantom hole information to allow accurate characterisation of the data before computer plotting. In addition to the usual two dimensional plots of geological or coal quality parameters, computer graphics has now been developed to produce 3-0 perspective borehole plots (with lithology) for correlation by appropriate fence sections. Due to the overall layered character of the seams and the inter-seam strata, correlation by closure around a "loopH of sections is well suited to coal-bearing sequences. Such a correlation provides a check on the degree of understanding that has been attained with respect to the geology of the area. At mine sites, a further refinement has been to construct clear perspex models of the mine, with

perspex uprights cut and placed over a base at locations to represent each highwall or every second highwall position. Down dip sections, as appropriate, are also fabricated from clear perspex. Horizontal and vertical scales of I :5000 and I :500 respectively control the model size, and at Goonyella, this meant a model 3 metres in length. Such a model could be assembled from three separate modules, each module representing Mapping based on a strike interval of 5km. oblique aerial photographs was copied, after scale adjustment, on to film with an adhesive backing and placed on the appropriate perspex uprights. With the successive highwall sections assembled , down dip correlations were carried out manually, aided by the results of down-hole photographic studies. The visual clarity offered by physical models of this type assist the geologist in dealing with problems of correlation, structural interpretation etc. at the mine site. However, such models have also been found to be more than useful to mine production and engineering staff in optimising pit layout and scheduling operations.

CONCLUDING SUMMARY ' properties of the coal measure Engineering sediments exposed in open strip mining operations present, under some circumstances, a serious threat to the economic or operational viability of coal extraction by this mining method . Depending on the expectations or experience in confronting such difficulties, a considerable amount of geological analysis of the sequence is often warranted. Important features of the sediments to be noted during investigations conducted from the surface include:

116 (i) the extent and character of superficial soil deposits and deposits of unconsolidated sediments; (ii) the base of the weathering zone appropriate to the design problem i.e. the degree of weathering that affects pit stability or the degree of weathering that affects excavatability; (iii) geological correlation of the dominant lithofacies making up the overburden and floor materials; (iv) any strata carrying disturbed or clay coated bedding surfaces, or shallow dipping j_oints with similarly deformed surfaces; (v) particular structural features including bedding plane shears, overthrusts and normal faults; (vi) separation of layering (bedding) and joints; (vii) the dip angle and dip direction of layered strata in the overburden and floor relative to the dip of the coal seam or seams as determined by core orientation or downhole techniques; (viii) as an extension of (vii) , the delineation of sedimentary folds within the overburden strata, particularly the centre or axis about which dip angle and dip direction vary. An index of the strength and sta bility of strata on exposure is useful if core is recovered . This can be obtained if the core is tested at the drill site using a Point Load Strength Index and simple slake testing devices. Where the sequence is exposed in mine highwalls, details of stratigraphic associations should be mapped as a basis for correlating down-dip information . Also of impo rtance is the measurement of the persistence of deformed or weak structural features and joi nts, which can also be projected down dip using these stratigraphic associations. Such information is basic to open strip mine design. The effort required to obtain it, bearing in mind the necessary manpower and drilling costs, may be considerable. Techniques have therefore been developed to allow rapid , yet accurate acquisition, analysis, processing and presentation of data in successive steps. By minimizing the amount of core required, and by developing field instrumentation that can be operated by nongeological staff, costs associated with acquiring the data are substantially reduced . Conversely, better drill hole coverage is possible for a given cost. Thus the evaluation of the large volumes of ground necessarily associated with coal mining is possible using modern techniques such as those currently under development.

REFERENCES Boyd G., Komdeur W., Richards B., 1978. Open strip pitwall instability at Goonyella mine causes and effects. Proc. Ann. Con/ Australas. Inst. Min. Metal/., No rth Queensland, pp 139 - 157.

Boyd G .L. , 1980. Structural diagenesis of coal measure sediments, M.Sc. Thesis, James Cook University of North Queensland, Townsvil/e.

Cox R.H .T ., 1979. Analysis of the stability of a highwall in an open cut strip coal mine. Confidential Report to Utah Development Co., CS/RO Division of Applied Geomechanics.

Emerson W.W. , 1967. A classification of soil aggregates based on their coherence in water. Australian Journal of Soil Research, Vol. 5, No.I, pp. 47-58.

Godfrey .H., 1978. Highwall stability in Goonyella open pit mine with particular reference to clay types, M.Sc. Thesis, James Cook University of North Queensland, Townsville.

Harries G., 1973. A mathematical model of cratering and blasting. Australian Geomechanics Society National Symposium on Rock Fragmentation, Adelaide.

Lazlett G. M., 1980. Statistical analysis of high wall and spoil pile failure data - Goonyella mine. Confidential Report No. 49 to Utah Development Co., CS/RO Division of Applied Geomechanics.

Mallet C.W. , Flood P.G. , Nguyen V.V., Ledger P . 1980. Depositional environments in the Baralaba Coal Measures - Moura, Queensland. Geomechanics of Coal Mining Report No. /9, CS/RO Division of Applied Geomechanics.

Richards B.G., Worswick M ., and Torok P., 1977. Results of material testing - ramp 6N spoil pile and ramp IOS highwall, May, 1977. Confidential Report No. 32 to Utah Development Co., CS/RO Division of Applied Geomechanics.

117

Techniques for the determination of geomechanical parameters during surface exploration J.R.Enever and R.M.McKean CS/RO Division of Applied Geom echanics, P.O.Box 54, MT. WAVERLEY. VIC., 3149

ABSTRACT A number of relatively simple techniques are currently under review by CSIRO Divisi on of Applied Geomechanics , to examine their potential for providing relevant information on geomechanical parameters during surface exploratory drilling. All of these techniques have been in existence in some form or another, but have not necessarily been widely applied to coal measure strata. The prime aim of the work being undertaken by CSIRO is to evaluate and further develop these techniques, with the eventual aim of encouraging their widespread use by exploration geologists and / or drill crews. The techniques under review fall into three broad categories: Techniques to assess geological structure: Automatic borehole camera; Borehole T.V. camera; Borehole impression packer; Techniques to determine mechanical properties: Point Load Testing; Borehole Penetrometer; Drillability Monitoring; Borehole Pressiometer; Borehole pressuremeter; Techniques to investigate stress fields: Open-hole hydraulic fracturing;

INTRODUCTION Geomechanical data can be defined as data describing the mechanical properties of a rock In coal mining, a knowledge of mass. geomechanical data is important in the assessment of aspects, such as the stability of highwalJs and spoil piles in strip mines, and the relative stability of long-term and short-term underground In order to be able to use this openings. information to best advantage, it is important for it to be available during the planning and development phase of a new project, and hence, should be obtained during the exploration phase. A great deal of information may be gathered from inspection of the rock mass in-situ, together with field and / or laboratory testing of core recovered from exploration holes. Although field and / or laboratory testing are common sources of information, there are some drawbacks resulting from inadequate drilling procedures, changes to core during transit and storage and / or limited core In other instances, the information drilling. required may not be obtainable from core at all (e.g. stress measurements). There is thus a need to develop in-situ techniques, capable of obtaining the information required . Geophysical techniques applied from the surface or from boreholes may yield much useful information, but work at the CSIRO Division of Applied Geomechanics is concentrated on non-geophysical borehole techniques. These techniques can be grouped into

categories, viz: (a) Geological inspection techniques (i) borehole camera (ii) closed circuit T.V. camera (iii) borehole impression device. (b) Techniques for determination of mechanica l properties: (i) drillability (strength) (ii) penetrometry (strength) (iii) pressuremeter (deformability) (iv) pressiometer (deformability) (v) point load tests. (c) Techniques to determine the environmental parameters of the rock mass. (i) hydraulic fracturing (stress) (ii) hydrological testing (hydrology) Although an interest is being maintained in all the approaches, effort is currently being concentrated on the evaluation of a few techniques for appli~tion in Australian coal basins [particularly numbers a(i), a(iii), b(i) , b(ii), b(v) and c(i)].

GEOLOGICAL INSPECTION TECHNIQUES Borehole camera Borehole cameras have existed in various forms for several years (Burwell and Nesbitt, 1954). Borehole photography provides a limited substitute for direct vision. To some observers, photography has greater meaning and value than

118

down-hole geophysical logging records. In many cases, the information from borehole photographs may be more reliable than that from recovered core (for example, the walls of holes may suffer less drilling-induced damage than core, particularly in soft rock). In the past, borehole photographs have generally not been oriented, and only qualitative or semiquantitative assessments could be made regarding the features identified. While this may suffice for some applications, it is generally unsatisfactory when attempting to assess the mechanical properties of a rock mass. A camera is being developed by the authors to take photogrammetric images over the entire length of 120mm to 300mm diameter holes in one run (overlapping consecutive photographs), and to allow quantitative analysis of the spatial orientation of features, such as bedding and joints, that influence rock mass behaviour. This is thought to be a significant improvement over current core orientation techniques. The camera has been designed to be used as a routine survey tool to supplement coring

(especially in areas of poor core recovery), and to reduce dependence on cores. Used in this fashion, it has the potential to improve significantly the cost-effectiveness of a drilling program, by improving the quality and quantity of the information obtained, and / or reducing the number of cored holes that need to be drilled . The effectiveness of the camera may be improved by using it in conjunction with closed circuit television (CCTV). CCTV could be used as a scanning tool to delineate areas of particular interest for subsequent detailed inspection by the borehole camera, which has much higher resolution. Development of the present design is based on experience with a prototype camera which used a conical reflecting surface to take photographs over a length of wall approximately 0.3 times the borehole diameter. This limited coverage was a drawback in the prototype. In the current design the conical reflecting surface has been replaced by a spherical mirror with hole wall coverage increased to 0. 7 times the borehole diameter accompanied by an increased, but quantifiable, amount of distortion. Figure I shows the optical configuration of the present camera. The space between the mirror and camera lens is open, to overcome problems D

Tangent & = At 6311

AB' Measured from photograph Figure 1

Detailed view of optical configuration of camera. 1. Spherical section, front surfaced, armour coated mirror, showing central hole through which the camera records on each frame, the depth and orientation data from a digital display. 2. Flash light window with compound curved surface designed to distribute light evenly over the viewed surface of the borehole wall. NOTE: the three legs joining 1 and 2. The larger of the two front legs carries ten electrical leads for power and data transfer between the two modules.

119 encountered with smearing of the glass window of totally enclosed cameras. Heating elements are incorporated in the lens and mirror surrounds to ensure freedom from condensation. The mirror is mounted coaxially with the optical axis of the Jens and has a central hole to allow display of orientation data for each frame photographed. This data includes the tilt of the camera from vertical, the tilt direction relative to an internal reference, the azimuth of the reference, and the depth from the hole collar. Figure 2 shows a typical borehole photograph with the relevant information displayed .

from the orientation transducers can be displayed on the SCU, along with a readout of depth obtained from the sheave wheel of the surface winch. The camera also includes warning devices to indicate end-of-film and camera immersion. Figure 3 shows the assembled camera and SCU. The centralising springs at each end of the camera are adjustable to suit hole sizes between 120mm and 300mm diameter. The electro-mechanical seven-core suspension cable has a breaking strength of 4.2 t, sufficient to ensure camera recovery from the hole. The camera module, behind the lens and ring flash window, contains the flash generation circuitry and exposure feedback control. Photo-interpretation is carried out by overlying photo images on a set of geometric grids corresponding to hole diameters between the limits specified. This system is designed to allow on-site analysis of spatial orientation data from the photos and recorded orientation data using a pocket calculator. This system should permit the dip direction and dip of features to be determined to approximately ± Z'. A number of bench trials have been conducted on the camera yielding the following specifications. Centrality in hole: typically ± 3% of hole rad i us for holes inclined up to 30" from vertical Depth measurement: 99.9m ± 0. 1m Azimuth measurement: ± 1° Inclination measurement: ± 1°

Borehole impression device

Figure 2

Typical photo from camera

Operation of the camera is essentially that of a 35mm flash camera, with focus and aperture preset for the nominal hole diameter. There is no shutter, exposure timing being achieved by controlling the flash duration through a photodiode feed-back circuit. For conventional speed films, flash duration is short enough to allow exposure "on the move". The exposure may be initiated manually, or automatically by programming the Surface Control Unit (SCU) to initiate exposure at chosen intervals of depths between preset lower and upper limits. The film magazine contains 300 frames, each of 28mm diameter, with a frame to frame cycle time of 1.4 seconds limited by the film wind-on speed. This means that for complete hole coverage the maximum traversing speed is 0.5D mm / sec, where D is the hole diameter measured in millimetres. Photography is always done on the retrieve (i.e. ascending from the base of the hole). The operation of the camera is controlled by two microprocessors, one in the camera and one in the SCU. The microprocessor in the camera is contained in a module behind the mirror, which also incorporates orientation transducers (digital compass, Haber and Rigby, 1975 and electrolytic tiltmeter, Enever and others, 1979). Information

As with borehole _cameras, borehole impression devices have been developed commercially over the past years (Harper and Hinds, 1977). In essence, these devices force a pressure sensitive "film" (e.g. wax impregnated paper) against a segment of the wall of the hole, leaving a permanent impression on the film. When recovered from the hole, the film represents a negative facsimile of the texture of the wall, showing features such as bedding planes, fractures, grain size, caving, etc. The practical advantage of this system over optical techniques such as T.V., introscopes and borehole cameras lies in its inherent ability to operate in muddy waters and mud columns, which would render optical systems inoperative . Because problems may be encountered with interpretation of the data obtained from impression devices, optical techniques are generally preferable, , but the obvious benefits of a system immune to optical interference provide the impetus for continued development. Effort is being directed primarily towards overcoming problems of interpretation. To date, experimental data has been obtained by employing a commercially available borehole impression device• to which the only major modification has been replacement of the standard orientation device with a remote reading magnetic flux gate orientation system to provide more reliable

*('I'riefus Industries (Aust.) Pty Ltd, borehole impression packer).

120

Figure 3

Borehole photogrammetric camera and surface control unit 1. Spherical section mirror (see Figure 1) 2. Flash light window and lens face glass (see Figure 1) 3. Orientation and micro processor control module 4. Camera module 5a,b. Top and bottom centralising springs 6. Electromechanical suspension cable

information. Orientation of the device is required to establish the spatial orientation of the features recorded on the film . An advantage of the remote reading system is that the device may be oriented down the hole in any preferred direction. This offers the operator the ability to orient the device preferentially, in the light of local geology, to optimise the record obtained . Figure 4 shows the complete system used for the work to date. The device consists of a steel inner tube which has an air or water expandable rubber tube fitted over and tightly clamped to it. Outside this are fitted two halves of a larger diameter steel tube, which has been split lengthwise. Each steel "split" is faced with sponge rubber and is in turn covered with a strip of replaceable soft, paraffin wax based film, which receives the impression of a l.5metre length of hole wall over a 100" sector in a 75mm diameter hole. The device is suspended by a steel cable or drill rods attached to the inner steel tube. At the base of the instrument is the orientation module which relates the azimuth of the impression film strips to magnetic north. In operation, film strips are laid over the rubber pads and held in place by a thin smear of grease along each edge. The device is lowered down the hole to the required position, and compressed air or water is supplied via the drill rods (or a separate hose if cable suspended) at a maximum pressure of 1200 kPa. Air or water enters the inner steel tube and exits via two bleed holes into the annulus

between the rubber sleeve and the steel tube. The rubber sleeve expands, forcing the sponge rubber faced splits against the wall. The impression is formed immediately, and the orientaion is noted. The pressure is then vented to atmosphere and with the collapse of the rubber tube the splits are retracted from the walls by rubber bands. The device is then withdrawn from the hole and the azimuth recorded on each strip. The device has been used in several holes, and has been found to function reliably both above and below the water table, even when some strata were water sensitive. Fine detail is usually quite evident. Figure 5a is a photo of a section of a typical record, taken under oblique lighting. This lighting markedly improves the contrast and highlights features for subsequent interpretation. This record was obtained from a hole drilled in steeply dipping Silurian sediments near Melbourne. A number of features seem unambiguous (bedding planes, high angle features), and these are included on the overlay interpretation shown in Figure 5b. There is, however, extra detail in Figure 5a which may represent genuine features of the hole wall or may be due to the experimental procedure. Definite identification of all features recorded is thus difficult, and has been a problem with the records obtained to date. Under present circumstances, identification is somewhat subjective. To make the process more objective, an experimental program is currently under way to establish

121

Figure 4

Borehole impression device 1. Metric scale 0.lmetre/division 2. Commercial impression device, with plastic impression strip in place on lower half-shell of unit 3. Connecting collar between device and flux gate orientation module 4. Flux gate orientation module developed by CS/RO 5. Surface readout ('slave unit") of orientation module 6. Borehole impression strip after use. (see detail Figure 5)

"control records" of various classes of features that can then be used to identify features obtained from field records. Identification of features would be aided by obtaining impressions of the complete circumference of any particular hole wall interval, as distinct from the partial records obtained to date. This may be achieved by multiple testjng at the same horizon. Alternatively full or near full circumference impressions may be developed in future instruments. Having clearly identified the trace of a feature of interest (Figure 5b), it is possible to establish its spatial orientation by measuring the coordinates of at least three points on the trace (preferably at 120" spacing round the circumference) relative to a set of cartesian axes. One axis corresponds to the hole axis, the other two axes having known rotational orientations with respects to the measured azimuth previously recorded on the impression strips (Dakhnov, 1959). Consideration is currently being given to the possibility of digitizing and computerising the analysis procedure. Apart from the major effort being devoted to analysis, attention is being paid to improving the hardware, including increasing the circumferential coverage, and the strip record length. •

TECHNIQUES FOR DETERMINATION OF MECHANICAL PROPERTIES Drillability studies Exploration programs ususally involve drilling holes. If these holes are cored, the core may be tested in the laboratory .or on site to determine material properties such as strength deformability, water sensitivity, etc. One disadvantage of this method, is that the results obtained may not be representative of the material in-situ. Weak portions of the sequences may yield core unsuitable for testing, biasing the results of a test program toward the stronger portions of the

sequence. Changes in conditions, such as water content or stress state, between the core and the site from which it was recovered may seriously affect the applicability of results obtained from core testing. For this reason there is merit in developing techniques for determining material properties in-situ in exploration boreholes. Apart from this, however, there is also the need to develop such techniques for operation in noncored holes which form a large part of many exploration programs. In-situ techniques may also be considerably cheaper and quicker to perform than corresponding laboratory procedures. Extended in-situ testing programs may therefore give a better appreciation of the situation for an area than tests conducted on limited core. The most direct and readily obtainable source of information on in-situ material properties is the drilling performance record . Fundamental studies of the correlation of drillability with rock strength have been reported by many authors, and a summary up until 1966 is given by Maurer ( 1966). The authors of the present paper are currently obtaining field data from a range of situations representative of Australian coal basins, .to establish to what extent existing theories that relate drillability characteristics to mechanical properties can be applied to Australian coal basins. A Longyear model 24 drilling rig is being equipped with instruments to measure penetration rate, thrust, drill stem torque, rotational speed and the pressure and flow rate of drilling water. Attention to date has been concentrated on continuous monitoring of thrust and penetration rate, but, information has also been recorded manually on drilling water pressure and bit rotational speed. Eventually it is planned to record all the above functions on the same time base automatically. Figure 6 shows the drilling rig with its current level of instrumentation. Penetration rate is monitored by means of a pair of jockey wheels riding on one

122

(a)

Figure 5

(b)

Impression device record. Three digit numbers give depth in metres 1,2 & 3. The up-dip 'noses' of planar fractures 4. Non-impressed area represents a hole in the wall of unknown depth 5. Possible bedding plane trace

of the hydraulic drive rods. The wheels are connected to a 10 turn potentiometer, producing an analogue output proportional to chuck movement during any one run (610mm). Thrust is monitored via a pressure transducer attached to the hydraulic system. Allowance is made for the self weight of rods. The electrical outputs from both sources are recorded on a two-channel chart recorder. Figure 7 shows a typical section of record obtained for a c<tred BX diamond drill hole drilled on Wambo Colliery Holding in the Warkworth district of .S.W. The left hand record shows

drive rod thrust for three consecutive runs. The disruptions to the record at the beginning of each run represent the periods during which steady drilling conditions were established. At present these sections of records must be ignored for purposes of quantitive interpretation, because they represent transient situations not readily amenable to analysis. This is a decided drawback with the present drill, as up to 20% of a run may effectively by excluded from interpretation. Apart from this, drive rod thrust appears to remain approximately constant for substantial increments, making it feasible to attempt quantitative interpretation.

123 DR IV E ROD THRUST 1kg) 1500

1000 500

PENETRAT ION (t. .J 600 400 200

24-6

25·0

Figure 7

Figure

Longyear model 24 drill rig with penetration rate and thrust instrumentation

The right hand record shows the corresponding penetration. The annotation on the chart indicates a number of changes in penetration rate, not all of which can necessarily be related to changes in thrust or lithology, as shown on the accompanying stratigraphic column. These changes presumably reflect variations in other parameters influencing drilling. There is an example in Figure 7 of the potential effect on penetration rate of a change in lithology in this case at a coal horizon within a

Typical section of record obtained with Longyear drill rig on Wamb o Colliery Holding

A marked increase in laminated sequence. penetration rate occurred in the coal at approximately constant thrust, possibly reflecting a change in mechanical properties. The apparent offset of the records in this instance is probably due to errors in Jogging procedures, which should be readily overcome by more attention to detail. Figure 8 shows a sample of averaged incremental plotting of penetration rate and thrust, extracted from the cumulative records. The results have been treated on the basis of two different criteria: firstly by averaging the penetration rate over increments corresponding to stratigraphic units identified in the core, and secondly, by taking the average rate for increments in which the rate appeared to remain approximately constant. In the latter instance, no knowledge of the stratigraphy was assumed . This corresponds to the situation that would pertain in open hole drilling. It is clear that there is a degree of definition in the data for this case that does not correspond with stratigraphy, and which, if the core had not been available, may have Jed to an interpretation of unwarranted and possibly The thrust, including rod misleading detail.

124 AVGE. INCREMENTAL AVGE. INCREMENTAL PfNURATION RATE I STRATIGRAPHIC I THRUST 1-/aiol I ,nctucing rod •oightl l kN I 0 5 10

DBl:RIPTIOII

LOG

(Ni - l•lrwted/Stl.iley

-~ SIi.al• • ffndt-G/C.,:bontcCoYI

Ht s.;:;t~~:·

.....

- - - - --

~ ~

Model !i i I

Drill data only

-->:"·

:..::-::·1---+-- - .......

♦~

L.-Model {i i : Drill data • core log s

:-·

:~ . . - - - + - - - - -

::-=;: Figure 8

Incremental penetration rate and thrust for p ortion of hole drilled on Wambo Colliery Holding

weight , remained approximately constant throughout the section shown . Variations in thrust alone would, therefore, not appear to explain the marked variations in penetration ra~e described above. These marked changes in penetration rate presumably reflect variation in parameters other than thrust influencing the drilling action. It is important that these parameters are identified and taken into account in future monitoring, particularly in the case of open hole drilling. In the case where the analysis was based on a knowledge of stratigraphy (cored hole) the variations in penetration have been somewhat evened out. This data most likely reflects variations in penetration rate related to the "average" properties of the stratigraphic units rather than localised or sporadic influences. It appears from Figure 8 that ~tratigraphic un_its containing significant proporttons of mud-size fractions have consistently lower penetration rates than the sandstone sections. Coal appears to fit somewhere between shale and sandstone in penetration rate. Laboratory unconfined compression strength (UCS) tests on samples of the shale and sandstone units shown in Figure 8 indicated similar strengths. This implies that the varation in penetration rate is not primarily a function of rock strength. It has been suggested (Maurer, 1966) that with diamond drilling, grain size has a very significant influence on penetration rate. Slower rates typically encountered when

drilling fine grained materials may be due to the effect of drill cuttings "clogging" up the areas between diamonds, thus inhibiting the effective transfer of thrust to the cutting edges. In the case of sandstone it is suggested that flushing water removes the cuttings to allow effective penetration. Figure 9 shows a more detailed penetration rate log compiled for smaller increments of the relatively uniform sandstone loca~ed at the bas~ ~f the log shown in Figure 8. At this level of detail 1t can be seen that there are quite dramatic localised variations in penetration rate. It is suggested ~hat these reflect variations in the efficiency of cuttings removal, effected by changes i~ water flushing conditions, rather than changes in rock strength properties. It is further suggested that the upper median value of penetration rate noted (approximately 50mm / min) rep~esent~ a c~n?ition in which the bit was "free" cutting with minimum influence of entrapped cuttings on penetration trate. On the basis of this assumption, it is possible to make an approximate estimate of th~ strengt~ ~f the unit from the recorded information. If 1f 1s considered that, for each revolution of the bit, the face diamonds cut into the rock by average distance equivalent to the average penetration rate / revolution, and that correspondingly the total thrust is equally distributed to the face diamonds only (no contact on matrix), then it is possible to relate the penetration rate and thrust information to the U. C. S. of the rock by means of the following expression (Enever and others, 1980a). Fi X k = u.c.s. (di) 2 where Fi is the average thrust / face diamond (equal to total thrust divided by number of face diamonds) di is the depth of penetration per revolution (equal to penetration rate / minute divided by rotational speed) and k is a factor depending on the chip formation characteristic s of the diamond / rock system . In the present instance, the average total thrust recorded was approximately IOk , including the influence of rod weight. This figure does not take into account the possible influence of water back pressure acting on the bit. The water pressure at the top of the hole was noted throughout the investigation, but no attempt was made to interpret this in terms of the down-hole back pressure. There were 280 face diamonds, and a representative penetration rate of 50mm/ min was assumed. The rotational speed was checked throughout and was found to remain approximately constant at 400rpm. The value of k was taken as 0.035, based on experience gained with a borehole penetrometer (Enever and others, 1980a). This represented a mean value between upper and lower limits determined for k for conditions approximately those in the current study. This gives an estimated value for the U.C.S . of the sandstone of 66.5 MPa. This can be compared with an average, laboratory-determ ined value of 43 MPa. The fact that the estimated value was higher than the laboratory value probably stems from an over-estimation of the average thrust per

125 PENETRATION l am / n11n l 0204060

;..-_· :: -~

.,.

:>. ·· :-- · - -' 1-- - - - - - - ~

•,

.. - - - - - - - -·

upper 111ed1an value assumed to correspond to "tru" cutting

Figure 9

Detailed penetration rate log f or sandstone shown in Figure 8

diamond . Factors such as the effect of water back pressure, wall friction , and partial clogging of the bit face with cuttings could cause the average thrust to be lower than assumed . In future it is intended to experiment with other types of drilling, in addition to diamond drilling, and to gain experience with a wider range of parameters in order to obtain a system capable of reliably differentiating rock types on the basis of fundamental strength properties.

This device has found (Kidybinski, 1979). widespread use, in operating Polish mines, in making semi-empirical predictions of likely roof caving conditions in extraction blocks. This is based on extrapolation from adjacent workings where roof caving conditions are known , employing relative strength profiles of the roof strata determined using the penetrometer. Considerable experience has been built up using this instrument, and the only apparent drawback is the need for a large diameter hole, which presumably has to be drilled using equipment not available in normal mining operations. A prototype penetrometer, capable of being used from underground in holes drilled by roof bolting equipment commonly used in Australia, has been developed and is currently being evaluated by the authors. It is envisaged that this will be a step towards the development of a surface-operating borehole penetrometer for use in exploration holes. Figure 10 shows the penetrometer "head" in detail, while Figure 11 shows the total equipment "package". The penetrometer head is located at the required depth in the test hole by means of the attached hydraulic hose. Interchangeable "backing shoes" cope with holes from 30mm to 45mm diameter (larger holes could be allowed for by the use of larger "shoes"). The hydraulic hand pump is used to force out the plunger, while the pressure gauge indicates, by means of a calibration factor (hydraulic pressure to plunger force) the plunger force at any stage. Plunger penetration is sensed

Borehole penetometry Although the direct use of information obtained during exploration drilling may potentially be the most convenient way of characterising rock strength in-situ, there are obvious problems that must be overcome in defining and reliably measuring parameters relating drillability to fundamental rock properties. The related concept of using borehole penetrometers in pre-existing holes avoids some of these problems, since the factors directly influencing performance are readily definable and can be accurately and fairly simply measured. In borehole penetrometry a plunger is pushed into the wall of a hole and the force-displacement curve recorded . The curve so obtained can either be used directly to given an indication of the relative strengths of various horizons tested , or it may be used to deduce quantitative rock strength. A borehole penetrometer operated in an exploration hole could be used to establish a "semicontinuous" profile of in-situ rock strength (c.f. the situation for limited core sampling and laboratory testing), either for the full length of hole, or at least for those critical portions of the hole above and below a coal seam intended for extraction. The Central Mining Research Institute, Poland , . has developed a borehole penetrometer for use in 86mm diameter holes drilled up to 15m into a roof sequence from an existing underground opening

Figure JO

Roof-bolt hole penetrometer head

Figure 11

Penetrometer and ancillary equipment: control box with hydraulic pressure gauge and displacement indicator

126 by the potentiometric transducer and reads out directly in millimetres at the control box. The plunger has a maximum extension of approximately 10mm and a maximum working force of approximately 2t. Field trials at Ellalong and Nattai North Collieries have been reported by Enever and others ( 1980a). At Ellalong, the practical viability of the device to operate in roof bolt holes was established . At attai orth the results obtained were evaluated , both in terms of their immediate qualitative usefulness and their ability to interpret fundamental rock strength information. The interpreted values of strength were compared with the results of laboratory tests conducted on corresponding core recovered from diamond drill holes adjacent to the test holes . Figure 12 shows representative force-displacement curves, obtained with a flat-ended tip in the penetrometer plunger, when testing a hole containing two basic rock types (a relatively "soft",coarse grained sandstone and a much "harder", more "brittle", fine grained mudstone). The tests conducted in the sandstone all yielded curves typifying "punching shear failure" at a maximum load considerably less that the capacity of the device. The mudstone could not be made to fail within the operating range of the device. The immediate qualitative value of the device, in this instance, is demonstrated by the comparison, shown in Figure 13, between a profile established on the basis of the penetrometer tests and the corresponding stratigraphic column . Quantitative interpretation of the results of the tests conducted in the sandstone was made by assuming that the tests approximated the bearing failure of a cohesive material (Wilun and Starzewski, 1975). The average value of cohesion so obtained (assuming the value of the angle of friction determined from laboratory tests) was 4.5 MPa, compared with the equivalent laboratory value of 5.7 MPa. Figure 14 shows typical force-displacement curves, obtained with a pointed tip, when testing a hole containing various sub-types of generally hard , brittle, rocks. These curves all reflect a failure mechanism comprising brittle chip formation interspread with crushing under the tip. This mechanism can be compared with the postulated model of diamond bit penetration proposed above. In Figure 14, the conditions existing at chip formation in each instance may be considered as a scaled-up version of those applying to each diamond in the drilling analogy. Using the same general analysis (Enever and others, 1980a) as outlined previously, the U.C.S. of each horozon tested was deduced from the tip force and corresponding tip penetration at chip formation . The results of this analysis are shown in Figure 15, displayed for comparison with the stratigraphic column. Figure 15 illustrates the qualitative value of the results when differentiating between stratigraphic units. Quantitatively, the average upper bound value of estimated strength determined for each stratigraphic unit appears to show reasonable agreement with the corresponding laboratory

"' ~ - - - - - - - - - - - - - - - - ---, .-'-

coarse sandstone

:;:

0..

;:::

TI P PENETRATION Imm )

Figure 12

Typical force-displacement curves obtained using penetrometer with a flat-ended tip

E r-

6089

-0

i..n

E _- -.,. --:to.

IT'I

<>· • •

T""""

. .....

Roof o ·

0 0·2 0·4 0·6 0·B 1·0 1·2 1·4 1·6 PENETRATION FORCE (t) Figure /3

Summary of penetrometer results obtained at Nattai North Colliery using flat-ended tip, showing comparison with stratigraphic column

127

~~ - - - - - - - - - - - - - - - - - - shale 0

:;:.,:...

2 l TIP PENETRATION {mm )

Typical force-displacement curves obtained using penetrometer with a pointed tip

Figure 14

_,E

Individual values based on high estimate of

average --- -- - average

.. low

average laboratory det ermined LI .LS. - normal to be:dd ing

)

average

parallel "

~ -='---'2~0_ _4...L0--6.LO--BLO- - ,--'-oo,-------,1-",20-----,14'="0ESTIMATE □

Figure 15

U.C.S.

---,-,J160

I MPa)

Summary of penetrometer results obtained at Nattai North Colliery using pointed tip, showing comparison with stratigraphic column

determined strength (normal to bedding for sandstone and laminated material, parallel to bedding for shale - see Figure 15). The upper and lower bound values of the strength estimates correspond to the maximum and minimum values of the factor k respectively (see section above on "drillability studies"), determined in the laboratory for conditions relevant to the field testing program.

TECHNIQUES TO DETERMINE ENVIRONMENT AL PARAMETERS Stress measurement by hydraulic fracturing Information on in-situ stress can be of assistance in planning mine layouts if that information can be obtained from the surface prior to commencement of development. By the time stress measurements can be made from underground (typically 2 to 3 years after the development decision), the data may be too late to influence major planning decisions. There is a valuable role for underground stress measurement in specific

aspects of detailed planning (e.g. roadway sizes, support requirements, panel layouts). However, decisions on such things as the best sites for major installations (shafts, drifts) and overall directional layout within a holding, require at least a preliminary estimate of the in-situ stress field throughout the lease to be made from the surface, if information on stress is to be of any value. Several techniques exist that are capable of measuring rock stress in relatively short holes (10 to 20m) drilled from the surface or from underground openings. There are only a limited number of techniques, however, that show the practical potential to be used in deeper exploration holes (Enever, 1975). Of these, hydraulic fracturing appears to be a relatively simple technique, amenable to use as an adjunct to routine exploration drilling, and with an extensive body of research to substantiate it. Considerable effort has been devoted worldwide to the application of hydraulic fracturing, in estimating stress in massive crystalline rocks, and in the deeply buried sedimentary sequences typically encountered in petroleum engineering. The authors are currently investigating the application of hydraulic fracturing in stratified coal measure sediments, at depths typical of Australian coal basins (30m to IOOOm). In hydraulic fracturing, a section of a hole is sealed off and pressurised at a steady rate until a fracture is initiated in the wall of the hole. For a vertical hole in a homogeneous medium, a vertical crack would normally develop in the direction of the major secondary principal stress component in the horizontal plane. The orientation of the induced crack can therefore be used to determine the directional arrangement of the stress field in the horizontal plane, while information obtained from the pressure record of the test (crack initiation pressure, crack shut-in pressure, crack propagation pressure) can be used to estimate the magnitudes of the horizontal secondary principal stress components (Enever, I 979). For application in coal bearing strata, the following potential problems were identified from the initial laboratory work: (a) the effect of bedding plane weakness, common in coal measure sediments, which results in the development of horizontal cracks (from which only information relating to the vertical stress can be obtained) rather than vertical cracks, (from which information relating to the horizontal stress field can be determined); (b) the effect of near vertical, incipient weakness planes (such as weakly cemented joints) intersecting test horizons, can cause crack development in a direction other than that dictated by the prevailing stress field. This limits the information obtained regarding the horizontal stress field; (c) the effect of test horizon permeability, which influences crack initiation and the subsequent interpretation of pressure records. Field trials have been conducted in short holes, drilled from the surface and from underground, to study the ability of the technique to give reasonable estimates of in-situ stress, before

128 developing the technology necessary for work in deep holes. The use of underground sites has meant that tests could be conducted at depths of up to 400m without the need for deep holes from the surface. All the above problems have presented themselves, to some degree, in the field trials. Operational and / or interpretative approaches have been developed either to overcome the problems substantially or to make the most use of the limited information available for some tests. For tests from the surface, a commercially available, inflatable, B size, straddle packer has been used , run on AW size diamond drill rods that have been modified by the addition of an "0" ring seal in the couplings to allow the rod string to be used as a high pressure hydraulic conduit. Rods are handled by the drill rig used to drill the test Hydraulic pressurisation is achieved by hole. means of an electrically powered , high pressure hydraulic pump (70 MPa). Continuous pressure recording is incorporated in all tests. Figure 16 shows the general experimental arrangement, including the hydraulic pump and the pressure recorder. This system has been used satisfactorily to a maximum depth of approximately 60m, and Figure 17 shows a typical pressure record Figure 18 shows the corresponding obtained. impression packer image of the vertical crack developed in this particular test (Enever and others, 1980b). For underground work, a compact, portable (power independent) set of equipment has been built up . The test packer employs commercially available inflatable sealing elements to operate in a B size hole. The underground equipment shown in Figures I 9 and 20 is manually handled and

Figure 16

operated , has been used in both vertical up and down holes to a maximum depth from the hole collar of approximately 20m, and has proved reliable and easy to use. Figure 20 illustrates the experimental arrangement employed underground, showing pumps and clockwork pressure recorder. Separate pumps are used to inflate the packers, and to pressurise the test interval. A typical pressure record obtained is shown in Figure 21. Underground tests have been conducted at a number of sites representing a range of geological environments. In all cases the opportunity has been taken to compare these results with the stress field measured by overcoring at the same site (Jaggar and Enever, 1978). In some cases, site conditions affected the hydraulic fracturing tests in such a way that only a limited amount of information could be obtained for comparison with overcoring results. Overall, there appears to be a fair correlation between the two techniques. The results have been progressively encouraging as more experience has been gained with the The results hydraulic fracturing technique. deduced for the site from which Figure 21 was obtained are summarised in Figure 22. In this instance the agreement in magnitude and direction between the secondary principal stresses in the horizontal plane determined by hydraulic fracturing (a 1, a 1) and the near horizontal principal stress components measured by overcoring is reasonable. Weak bedding has affected all field sites to some extent. In principle, this problem can be overcome by selecting a test horizon free of material with low bedding plane strength, but this may not be possible in many coal measure sequences. In the field work conducted to date from the surface, it

Arrangement for hydraulic fra cturing from surface.

Note hydraulic pump and pressure recorder

129 OOIENTATION CFVEflT(AL CRACK

ockcrlll'ltot,on

, • (t~rox= shut 1n pressure= 1..1irf2 1q-o; .sfr,ngth : crcxk rwt10t1on

assuming ; crock wut1ot10n - strtngth approx = crock propogot1on q' = l C( - crodq:ropogollon .a ~ 1f: J. f75- 24S

Figure

Typical pressure record fo r surface hydraulic f racturing obtained on Wambo Colliery Holding

has been possible to find suitable test horizons (massive sandstones) of adequate thickness (at least 1.5m for the geometry of the tool currently employed) at convenient locations for testing, within generally variable stratigraphic sequences. Given this condition, testing with the current equipment appears reasonably satisfactory, and interpretation relatively straightforward. Suitable test horizons can be readily selected by inspection of core recovered during test hole drilling. Initial attempts at underground field work were seriously hampered by the problem of weak bedding (particularly in conjunction with high Development of a more horizontal stress). compact and axially rigid equipment, shown in Figure 20, has greatly improved the chances of being able to select a suitable horizon for testing and obtaining a vertical fracture amenable to interpretation. This compact equipment has a test interval length of 0.3m · compared with 1.5m for the surface equipment and its increased axial stiffness tends to reduce axial stress concentrations One drawback to this during pressurisation. equipment is the tendency for cracks, initiated vertically, to rotate rapidly towards the horizontal bedding planes, particularly as the cracks extend to near t~e stress concentration zone adjacent to

Figure 18

Impression packer image of vertical crack obtained on Wamb o Colliery Holding

the ends of the sealing elements (Enever, 1979). Care must be taken to stop pumping immediately crack initiation occurs, so as to obtain a shut-in pressure representative of the horizontal stress components, before crack propagation induces rotation. This problem is not apparent with the longer test horizons associated with the surface equipment, presumably because of the greater degree of longitudinal crack extension required to bring the crack into the proximity of the seals. The problem of near vertical, incipient weakness planes intersecting test horizonts has been encountered to a limited extent both in underground and surface testing. There is little that can be done about this problem, except to try to avoid such intersections in the horizons selected for testing. However, the problem is not as serious as the effect of bedding, since, with the tendency to produce a vertical or near-vertical crack, some information is obtained on the magnitude of the It appears that the horizontal stress field. magnitude of the horizontal or sub-horizontal stress component normal to the plane of the

130

. '!} __, ► an I 1

Figure 19

Equipment used for underground hydraulic fracturing. From bottom: Impression packer; Testing packer; Installation rods and holes. The audio sensor and headphones are used to assist in detection of crack initiation

Figure 20

Experimental arrangement employed underground, showing pumps and clockwork pressure recorder

131

Figure 21

• , • ,

Tut Na.1 Test No. 2 Test Na. 3 Ttst Na. 4

Typical pressure record for underground hydraulic fra cturing obtained at Tahmor Colliery hydraulic fracturin

Hll'lt Grid North

secondary o r incipal

, - ----t-- ~

(a)

(b)

Figure 22

Comparison of overcoring and hydraulic fracturing results for Tahmor Colliery

weakness can be approximated by the shut-in pressure obtained from the test record. Although this does not help to specify the orientation or the relative magnitudes of the horizontal stress components, it does give some idea of the order of magnitude of the horizontal stress field relative to the vertical stress component, and this can be of considerable significance for planning. Considerable difficulty had been anticipated from fracture fluid penetrating into any test horizons with appreciable permeability (such as sandstones, which might otherwise be preferred for testing In practice, because of their homogeneity). however, it has been found that most horizons selected for testing appear to be effectively impermeable to the fracturing fluids initially used (hydraulic oil for up-holes underground and water Except in one instance, no for down-holes). significant problems have been encountered either with initiating cracks or interpreting pressure records. In one instance, it was necessary to use a viscous, water-based , fracturing fluiq (hydropol) in a very coarse grained porous sandstone to bring about crack initiation. More field testing in a variety of conditions is required to assess where additional developmental effort is needed. This is particularly true in relation to the effect of bedding, where it must be established, by experience, in what proportion of cases a vertical fracture cannot be initiated with the current technology, and whether or not it is worthwhile lo_oking into a more elaborate means of unsuring vertical crack initiation. Another aspect that requires further investigation is the question of determining a representative value of rock strength for interpretation of pressure records. Experience to date has shown that it is not always reliable to use a value determined by laboratory testing of the recovered There appears to be a tendency to core. overestimate the effective strength of test horizons by this means, resulting in a corresponding overestimation of major stress components magnitudes. There is evidence to suggest that the effective strength of a test horizon can be estimated directly from the test record, but this aspect needs further investigation (Enever and others, 1980b). A new fracturing tool for surface operation has been built (Figure 23) and is currently being evaluated. This tool combines all features thought desirable from experience to date. It has an adjustable test horizon length, that can be matched to the physical characteristics of the sedimentary sequence being tested (particularly the length of horizons free of weak bedding planes). Other features include simplified operation, making it easier to interpret pressure records, down-hole pressure transducing (to allow for the effect of hydrostatic head when testing in deep holes), and the capability of working in 75mm diameter holes, which are much more common than the 8 size hole required for the previous equipment. To facilitate the conduct of hydraulic fracturing in deep holes, a so-called nendless tubing unitn is being developed. In principle, this unit consists of a continuous steel tube wound on to a trailermounted, large diamet_er, drum. The tube can be

132

(a)

(b)

Figure 23

Modified fra cturing tool showing do wn-hole pressure monitoring and seal pressure ndumping" arrangement

133 used to lower or withdraw a tool from a test hole at relatively high speed (approximately 60m/min). The tube will also act as a high pressure (up to 40 MPa) conduit to supply the fluid for hydraulic fracturing. This will avoid the need for joining together rigid rod lengths, which is rime consuming, labour intensive, and may be unreliable. Figure 24 shows the first part of a multi-stage development. The basic operating functions of the eventual unit are contained in this prototype, which has been field tested and appears to operate satisfactorily. At present, a second stage of development is being undertaken to equip the unit with a number of features necessary to fulfil all likely requirements.

Figure 14

• Endless Tubing Unit" for deep hole hydraulic fracturing first stage development

REFERENCES Burwell, E.B. and esbitt, R.H ., 1954. The X borehole camera. A/ME Transactions. August, 1954. Mining Engineering pp. 805-808. Dakhnov, V.N., 1959. Translated by G.V.Keller, 1962. The application of geophysical methods; electric well logging. Quarterly of the Colorado School of Mines. Vol. 57, No.2. Enever, J.R., 1975. The measurement of rock stress at depth from small diameter surface drill holes. Proc. Aus. J.M.M. Ann. Conf., Adelaide. Enever, J.R., Tchepak, S., Skotnicki, A., 1979. Electrolytic transducers as monitoring devices for lateral earth movements in civil engineering works. Aust. Geomech. J. 42-67. Enever, J.R. , 1979. A laboratory investigation of hydraulic fracturing as a technique for estimating in situ rock stresses. 4th Tewkesbury Symp. on Fracture, Melbourne, February, 1979.

Enever, J.R., McKean, R.M., Crawford, G.R. and Lupton, G.T., 1980a. Application ofa roof bolt hole penetrometer to the determination of in situ strength of coal measure sediments. CS/RO Aust. Div. of Applied Geomechanics, Geomechanics Coal Mining Report No. 27, 26pp. Enever, J .R., Wooltorton, B.A., Crawford, G.R., Cunningham, D., and Dean, A.K., 1980b. Stress measurements on the Wambo Colliery Holding and their interpretation in terms of roof conditions. CS/RO Aust. Div. Applied Geomechanics, Geomechanics Coal Mining Report No . 28, /9pp. Haber, M.B., and Rigby, G.A., 1975. A digital compass using monolithic optoelectronics. Inst. Radio and Electronic Engineers Convention (/975) Digest of Tech. Papers. Harper, T.R. and Hinds, D.V., 1977. The impression packer: a tool for recovery of rock ma s fracture geometry. Proc. I st lnt. Symp. • Rocks tore '77" - Storage in excavated rock caverns, Stockholm, Sweden, Sept. /977, Vol. 2, pp. 259 -266. Jaggar, F.E., and Enever, J.R., 1978. An evaluation of different stress measuring techniques conducted in the southern development headings at Pelton Colliery. CS/RO/ ACIRL Geomechanics of Coal Mining Joint Report No . 8, 32 pp. Kidybinski, A., 1979. Experience with hard rock penetrometers u ed for mine rock stability predictions. Proc. International Society for Rock Mechanics, 4th Congress, Montreux, Switzerland, 2: 293-301. Maurer, W.C., I 966. The state of rock mechanics knowledge in drilling. Proc. 8th U.S. Rock Mech. Sym., Minn. 355-395. Wilun, Z., and Starzewski, K., 1975. Soil Mechanics in Foundation Engineering, 2nd edn. 2 vols. Surrey Univ. Press.

134

135

Engineering geology and the environment do coal geologists have environmental responsibilities? G.R.Mcllveen Geological Survey of N.S. W., Box 5288, G.P.O. SYDNEY, 2001.

ABSTRACT Environmental planning for the impacts of coal mining is as much part of mine planning as any other design component. Planning rehabilitation is also part of this planning process, and, therefore, part of the mine design, rather than a procedure which can be postponed till a later date. Environmental planning procedures should aim to achieve informed and realistic decisions, based on a coordinated evaluation of data on natural systems, resource demands, and likely mining developments. Geologists bring a natural science background to design teams otherwise composed largely of engineers. They can contribute at all stages of the planning process, including the exploration or design stage, the assessment stage, and the mining or monitoring stage. Basic geological data are needed for environmental planning, since the structure and composition of the resource determines the mining method adopted and hence the environmental impacts of its extraction. Collecting environmental data is done most cost effectively during the exploration program, providing the opportunity for effective design at an early stage. Other important environmental geological aspects are: hydrogeochemical parameters such as the relationship between the distribution of sulphur minerals and water quality; groundwater quality and aquifer characteristics and their subsequent behaviour as spoil; geotechnical aspects of long-term spoil and final high wall stabilities; and the heritage value of exposed geological features. Geological inputs at both site-specific and regional scales are also required at the stage when the project is being assessed prior to development approval.

INTRODUCTION The sub-title of this paper is "do coal geologists have environmental responsibilities?". This paper will argue that they do, and that by exercising them they can save money through increased efficiency in mine design, that they can save money through smoother transition through environmental planning controls, and that they are assisting the community through more efficient environmental planning and reduced environmental impacts.

WHAT IS ENVIRONMENTAL PLANNING? Environmental planning is an endeavour to take the natural environment into account in the overall balance sheet when planning economic development. Economics and the environment For example, economic cannot be divorced . development does not resolve land use conflicts on any more than short-term economic criteria. Nor does it adjust itself readily to natural systems and constraints as they really occur, rather than as they are perceived at the time of development. To choose an obvious example of this, consider housing development on flood-prone land. Comprehensive environmental planning takes into account: (a) the natural constraints of the physical environment;

the long term economic criteria, in particular the conservation of resources by, for example, avoiding sterilization of potential mineral resources by urban development; (c) the non-economic criteria, by comprehensive assessment of environmental impacts and by deliberate preservation of representative parts of the natural environment. The evaluation and resolution of physical constraints, long term economic criteria and noneconomic criteria, should normally be carried out Applying the in the planning of land use. principles of sound environmental planning to large scale development proposals, such as new coal mines, is really only an extension of existing It does not involve the planning techniques. development of new techniques, although it may require a more detailed approach ·in certain aspects, such, for example, as groundwater.

(b)

WHAT DO WE SEEK TO ACHIEVE WITH ENVIRONMENTAL PLANNING FOR COAL MINING? At a regional scale, successful environmental planning for new developments should: (a) examine possible land-use conflicts, such as those with intensive agriculture, and indicate ways of minimising them; (b) ensure the conservation of coal resources by avoiding their sterilization by other land uses; (c) assess the regional environmental _impacts

__I

136 likely to result from the proposed developments and propose guidelines for reducing or containing the adverse impacts of such developments. One can argue that these are really tasks for geologists in government, especially environmental geologists, and any others giving advice to planners. The point is that coal geologists tell the community where the coal is , so its extraction can be planned in an environmentally sound manner. On a local scale , the objectives which environmental planning seeks to achieve for a given mining project are: (a) to ensure that environmental factors are taken into account in the early planning stages of the project; (b) to anticipate and to minimise the environmental impacts of mining; (c) to ensure that the area of land disturbed by mining is returned to a desirable and productive use after mining; (d) to ensure conservation of the resource by ensuring maximum extraction and utilization of any given deposits; (e) to minimise the environmental impacts of the disposal of wastes after · coal preparation.

WHAT IS THE POLICY FRAMEWORK WITHIN WHICH ENVIRONMENTAL PLANNING FOR COAL MINING MUST BE CARRIED OUT? In N .S. W., a complete answer to this question requires a familiarity with at least the Coal Mining Act administered by the Department of Mineral Resources, the planning legislation administered by the Department of Environment and Planning, the pollution control legislation administered by the State Pollution Control Commission, and the Local Government Act. Environmental problems generally cut across the traditional boundaries between disciplines, and also across traditional administrative structures. Three steps in the development of a mineral deposit fit into this framework: exploration under some form of prospecting authority; application for a mining authority, for Crown coal at least, at which stage (in N.S.W.) an Environmental Impact Statement is required and lease conditions are set; and operational controls during mining and rehabilitation . In N .S. W., developers of prescribed developments, including mines, are required to submit an Environmental Impact Statement (EIS) describing in detail : (a) the proposed mining operation; (b) the existing environment; (c) interactions of the proposed operation with the environment; (d) the measures to be taken to control these interactions; (e) an overall assessment of the environmental impact. The EIS itself is reviewed by so many authorities having some right of "protection of interest" (such as the Department of Agriculture, the Water

Resources Commission, the Joint Coal Board, etc) that it may become not just an integral part of the approval process, but virtually the concrete manifestation of the whole planning and approval operation . Everyone gets to see it, from farmers with a local interest to the Commonwealth Government for export approval. This suggests that all its content, ranging from reserve estimates to estimates of increase in housing demand, will come under scrutiny by someone.

WHAT DETERMINES THE ENVIRONMENTAL IMPACTS THAT COAL MINING IS LIKELY TO HAVE? The immediate determinant is the mining process to be used. There are clearly fundamental differences between, for example, a strip mine and a multi-seam open pit, and of course, between surface and underground mining. The mining method to be used is, in turn, determined by the most efficient means available to win coal from a given geological setting. It is the internal nature, the overall form and the size of the deposit itself which determine the likely environmental impacts of mining that resource . An understanding of the basic design elements of the development is essential to any study of its environmental impact. Environmental planning for coal mining must be placed on a firm basis of realistic, factual data and experience; that is to say on the information collected and interpreted for engineers by coal geologists . The environmental investigations for any major coal mining development can be seen as being undertaken in three stages: (a) collection of background data, which will also indicate constraints on development. This data should be collected from the earliest time, i.e. during prospecting and the feasibility study; (b) the incorporation of environmental data into the design of the mining operation and the accompanying progressive rehabilitation; (c) the preparation and submission of these plans and an Environmental Impact Statement.

THE NEED FOR BASIC GEOLOGICAL DAT A IN ENVIRONMENT AL PLANNING FOR COAL MINING Coal geologists , engineering geologists, hydrogeologists and environmental geologists can undertake any aspect of these environmental investigations. At the exploration stage, they have firstly to see that that operation is itself being carried out with as little impact on the environment as is necessary. Added to this is the need, in the early stages of the project, to start collecting as much environmental data as is likely to be of importance in the subsequent stages. The sorts of basic data that are likely to be important are not listed here, because it is important to recognise that environmental geological data are not different data in any way from other geological data. They are either that basic data which should be used for environmental

137

planning, or, at most, an extension of that basic data to provide additional insights for mine design which will relate the design to the environment. Consider two examples of information which should be able to be derived from a drilling program with little extra cost other than that of laboratory testing and interpretation. A knowledge of rock properties gives insights into predicting blasting behaviour. It also gives data on overburden characteristics and groundwater conditions for rehabilitation , for re-establishment of aquifers, for short and • long-term water pollution control and for erosion control. Coal quality data ,n;: ,1ot only important for marketing, but also give information on likely mine water quality, washery reject characteristics, and the hazards of coal stockpiling. A careful appraisal of local lithological conditions for ground vibration studies will be of greater value if it is seen as part of the environmental design for the mine, rather than if one goes ahead with a purely engineering study and then is required to interpret the more limited data from such a study at some time in the future . Such an interpretation might be required , for example, when people living in the adjoining residential area start submitting claims for compensation , or even later when the company applies for another area adjoining a second village. Within a relatively short space of time, rehabilitation has come to be seen by the industry as an integral part of the process of winning coal rather than as a burdensome afterthought. Since it is essential that rehabilitation be designed to form an integral and continuing part of the mining operation, it is essential that it be planned at the same time as the operation is designed . A broad determination of how the land is to be used after mining should be made at the design stage in order to make the most efficient, cost-effective plans for such integrated rehabilitation. In N .S. W ., this could even include building development in the future . Rehabilitation is in fact only the preparation, at an acceptable cost, of the land for a desired use after mining. Efficient planning for rehabilitation in open cut mining depends on the determination , from the e xploration prognm , of many of the mineralogical , chemical , and physical properties of th e stratigraphic units that form the overburden and mid-burden materials. The aim of such studies is to arrive at predictions of the long-term behaviour of the overburden once it becomes regraded spoil. Geotechnical and soil science disciplines are needed to assess the long-term stability of the spoil with respect to erodibility, slippage settlement , and subsidence from future underground mining. These disciplines are also needed to assess the behaviour of particular buried materials such as washery rejects and any disagreeable lithologies such as highly saline or sulphur-rich rock materials. Other problems which need to be investigated are hydrogeological aspects such as the reestablishment of aquifers, the long-term future of the final void and associated high wall , and any potential leaching by water throughflow which

may eventually contribute to the quality of the surface waters. In the long term, we face the challenge of environmental impacts due to mining through or under riverine flood plains, and of open pits that are several orders of magnitude larger in volume than anything developed to date. The time to do the necessary work is before the second company applies for approval and the first development is cited by someone as a good reason for refusal. At this point one might consider that geologists may even be called on to look into the impacts of not mining a particular deposit. If one looks beyond the immediate mine area, the impacts of not mining may be less desirable than those of mining. For example: (a) if this deposit is not mined, where will the coal come from to fill the need? Perhaps it will be more desirable to mine this deposit than to suffer any of the alternatives; (b) on the other hand, if we do without the equivalent amount of coal from this deposit altogether, what will be the environmental impacts of doing without the energy produced , or the export earnings produced , or of obtaining the energy from another source? These questions require careful consideration, not only by geologists . This paper has concentrated on the environmental aspects of new open-cut coal developments. This is where there is at present the greatest room, and the greatest need, for geologists to contribute. Other aspects not covered in detail here include underground mining and subsidence problems, the government assessment aspect of environmental problems, or the role of the consultant geologist in environmental design teams. This paper has not looked back in time to see where we could have done things differently by having involved geologists more in the art of influencing engineers. The question of the heritage valtJe • of exposed geological features has not been considered . ow is the time for geologists to express clearly how important their advice is in planning for, and assessment of, any development which disturbs the environment. There undoubtedly are some large ones on the way.

ACKNOWLEDGEMENT This paper is published with the permission of the Secretary, .S . W. Department of Mineral Resources.

Typesetting & Compilation:

Printing:

CARTOSCOPE PTY LIMITED , 9th Floor, 77 Pacific Highway, ORTH SYD EY , .S.W. 2060 Ph 02 - 929 7431

ALPHA PRI TI G SPECIALISTS PTY LIMITED, 105 Charles St, ST MARYS , N.S .W. 2760 Ph. 02 - 623 1590