Australian Longwall Geomechanics - A Recent Study

9
AUSTRALIAN LONGWALL GEOMECHANICS - A RECENT STUDY BY Russell C. Fritb'. Ale. M. Sted & David Price? This paper presents recent findings concerning longwall caving mechanisms relating to Australian mining conditions and their effect upon longwall support requirements and behaviour. The findings are based on a program of intensive hydraulic support monitoring in a variety of Aushalii mining conditions coupled with geotechnical observations. Results from the study illustrate periodic weighting cycles of varying severity and these are dated to the overlying geology in terms of the stmgth of individual strata units, their thickness and proximity to the longwall extraction. Several different caving mechanisms are identified which contribute to periodic weighting cycles. The importance of rewgnising weighting cycles in relation to the operation of the longwall face will be discussed in detail. Longwall support design methods are critically examined in relation to the case histories of the monitoring program. The applicability of the said design methods to Australian coal mining conditions is discussed and an outline of design considesations to facilitate productive longwalling is presented. Overall, the paper only 'scratches the surface' of the work currently being undertaken by ACIRL in this area, and the findings of that work. 1 Principl Ground Support Consultant ~eolkhnical Engineering ACIRL LTD 2 Mining Engineer Geotechnical Engineering ACIRL LTD 3 Senior Mining Engineer Underground Mining ACIRL LTD INTRODUCIlON When discussing gwmechanical design techniques for longwall support specification one must fmtly consider the question "why do we mine coal using longwall methods?" The fOIem0St answer to this question nlates to high productivity and presumably profitability although safety issues can also be strongly argued in support of the use of longwalls. Given that productivity is the driving force behind longwall mining it must also be the major wnsideration in the gwmechanid design and specification of longwall supports. Reeat resarch by ACWL (sponsored by NERDDC) has monitored the performance of hydraulic longwaU supports on several Australian longwall faces in a variuy of mininglgwlogical conditions. This paper presents the initial findings of the research work and outlinc~ the. m h , ACIRL belieye should be applied to the gwmcchanical design of longwall faces. Given that the emphasis behind the g w d d design oflongwah is productivity, one must lmow which gwmechanical fac& affect longwall productivity. Two major factors have been identifidd: - FaceSpall - Roof Falls Face spall is not alway dcaimental to productivity as in some instances it can actually improve productivity by reducing sheam power requirements and allowing the sheanr haulage spced to be increased. Essentially, if face spa11 manifME iwelf as small to medium size pieces which the A.F.C. can easily transport then productivity will not navssarily be affected. In fact, some mines have noticad that their highest production occurs when the. wpporw are 1 lth International Ccnfennca on Grcund Conhol in Miming. The U n i i ol Wdlongong, t4S.w.. ~uly 1902.

Transcript of Australian Longwall Geomechanics - A Recent Study

Page 1: Australian Longwall Geomechanics - A Recent Study

AUSTRALIAN LONGWALL GEOMECHANICS - A RECENT STUDY

BY

Russell C. Fritb'. Ale. M. S t e d & David Price?

This paper presents recent findings concerning longwall caving mechanisms relating to Australian mining conditions and their effect upon longwall support requirements and behaviour. The findings are based on a program of intensive hydraulic support monitoring in a variety of Aushalii mining conditions coupled with geotechnical observations.

Results from the study illustrate periodic weighting cycles of varying severity and these are dated to the overlying geology in terms of the stmgth of individual strata units, their thickness and proximity to the longwall extraction. Several different caving mechanisms are identified which contribute to periodic weighting cycles. The importance of rewgnising weighting cycles in relation to the operation of the longwall face will be discussed in detail.

Longwall support design methods are critically examined in relation to the case histories of the monitoring program. The applicability of the said design methods to Australian coal mining conditions is discussed and an outline of design considesations to facilitate productive longwalling is presented.

Overall, the paper only 'scratches the surface' of the work currently being undertaken by ACIRL in this area, and the findings of that work.

1 Principl Ground Support Consultant ~eolkhnical Engineering ACIRL LTD

2 Mining Engineer Geotechnical Engineering ACIRL LTD

3 Senior Mining Engineer Underground Mining ACIRL LTD

INTRODUCIlON

When discussing gwmechanical design techniques for longwall support specification one must fmtly consider the question "why do we mine coal using longwall methods?" The fOIem0St answer to this question nlates to high productivity and presumably profitability although safety issues can also be strongly argued in support of the use of longwalls. Given that productivity is the driving force behind longwall mining it must also be the major wnsideration in the gwmechanid design and specification of longwall supports.

Reeat resarch by ACWL (sponsored by NERDDC) has monitored the performance of hydraulic longwaU supports on several Australian longwall faces in a variuy of mininglgwlogical conditions. This paper presents the initial findings of the research work and outlinc~ the. m h , ACIRL belieye should be applied to the gwmcchanical design of longwall faces.

Given that the emphasis behind the g w d d design oflongwah is productivity, one must lmow which gwmechanical fac& affect longwall productivity. Two major factors have been identifidd:

- FaceSpall

- Roof Falls

Face spall is not alway dcaimental to productivity as in some instances it can actually improve productivity by reducing sheam power requirements and allowing the sheanr haulage spced to be increased. Essentially, if face spa11 manifME iwelf as small to medium size pieces which the A.F.C. can easily transport then productivity will not navssarily be affected. In fact, some mines have noticad that their highest production occurs when the. wpporw are

1 lth International Ccnfennca on Grcund Conhol in Miming. The U n i i ol Wdlongong, t4S.w.. ~ u l y 1902.

Page 2: Australian Longwall Geomechanics - A Recent Study

Russell C. Frith, Alex M. Stewart & David Rice

in yield and the face spall is at a maximum. Under these wnditions the shearer supply acts as a power loader and shearer power is low.

If however, the face spa11 consists of largelvery large lumps which continually block the A.F.C. then productivity can be seriously reduced. Note also that large blocks of face spall pose a serious safety threat to face crews. Given that the majority of longwall faces observed suffer some degree of face spall, the next consideration is as to what effect will it have (if any) on the stability of the immediate roof?

Immedii roof stability beween the face and canopy tip has been observed to be a problem on many longwall faces, reducing productivity for the following reasons:

- large lumps can block A.F.C. or crusher.

- roof cavities make support setting difficult/impossible/time consuming

- large roof cavities requires cleaning and supporting

I lth lntematmsl Contenmce on Gmund Control in Mink

In addition, roof rock being mixed in with the R.O.M. Coal will in- the ash content and require added washing to produce a saleable product.

Roof stability is not only affected by face spa11 but also more importantly, the amount of roof convergence during the mining cycle. Essentially, high roof wnvergenm results in a heavily strwsed area (due to bending action see Figure 1) between the face line and the support. It is this mechanism which (in cwain strata conditions) is a maior cause of roof falls in fmnt of the longwall Iupports:

Given then, that face. spdl and immediate roof s t a b i are the major geomechanical problems asso5ated with longwall mining the question is posed as to 'What control do we have over their occumnce in tams of longwall specification and can we improve face. productivity as a result?"

More one wu, consider the productivity aspect of a partioular coal seam it is vital to understand (at least wnapually) the mechanisms that are causing both facc xpall and roof instability. Traditionally, all the arguments posed to explain face spa11 etc have been based on abutment stresses (se Figure 2) whgcby a sIms peak exists a short distance in fmnt of the facc Unc Dl. lk distance in front of the face line to the shgs peak is d&nnkd by the amount of face @I. Howeva, recent fidings by ACIRL suggest that whilst the s a J s abutment is undoubtedly t h m in order m arrive at a uniqw solution to each mining condition, one must consider the amount of roof or roof to floor C M I ~ ~ ~ ~ E C asso5ated with longwall mining. The real iuue in g m c s is not how st- redistribute due to mining but how the redistributed strcssea manifest themselves as convergences. This bdngs into question the role of local geology, srmcturr etc and this provides the basis for the dcsign approach that ACIRL Pre developing.

Convergence mtasurementsin geomechanicsare simple, acctuate an inexpensive with a short tum- around time, whereas strey m c a a ~ t s an complex, less accurate, expensive and with long turn- around timw required to calculate actual messes h m the mtasummts taken. Strata convergence therefom whilst pmvidiig a unique fingerprint for each mining

w, The Univenihl of Wollon00n0. N.S.W.. July 1692.

Page 3: Australian Longwall Geomechanics - A Recent Study

Russell C. Frith, Alex M. Stewart & David Prke

situation also provides easy measurement and control for the engineer. S a s s measurement is not however to be discounted as it pmvides valuable information in understanding the processes involved in causing strata convergences.

The starting point or fust consideration in the geomechanical design approach must be the WEIGHTING MECHANISM that the longwall face will experience. This will determine the caveability of the overlying strata and the magnitude of roof wnvergence me (i.e. loading) that the longwall will experience and variation thcnof during mining. Note that wnvergence m e (mmlhr) will be used in the discussions for the geomechanical aspects, as the weighting mechanism will determine the wnvergence rate. Only when coupled with the speed of mining can the absolute convergence affecting both face @I and roof stabiiity be assessed.

ACIRL have identified two wnceptua! weighting mechanisms from monitoring on Australian longwall faces:

1. VARIABLE WEIGHTING: conditions do not visibly change over protracted periods of monitoring and measured convergence rates remain within narrow bands. Figure 4 illustrates typical support behaviour in variable weighting conditions.

2. PERIODIC WEIGHTING: phases of light and heavy loading conditions both of which last for several consecutive shears. Vastly different magnitudes of convergence rate exist bet- the two conditions and visibly 'fferent mining mnditions. These two l c d n g mnditions are attributed to the m c e of a massive strata unit in close pmximity to the seam and its intermittent faulure. The period of light loading depends on length of cantilever required for massive strata to fail, heavy loading period determined by where failure aceurs ie in fmnt of face or at face line (see Figure 3). Figure 5 shows typical support behaviour in the two loading wnditions described. It should be noted that this effect has been noticed and described by other engineers. In fact Wilson [3] whilst rewgnising it, dismissess periodic weighting by arguing that over a long period of time the average convergences remain unaltered. Whilst this may be true, one must consider the short-term implications of periodic

weighting and the geomechanical problems it CBUm.

The pncire definition in engineering terms of what strata if massive is unclear at pmmt. It would sppepr though that t h e is some form of relatiomhip PC follows:

massive strata = f ( s m thicloless. lldistance from seam horizw)

In other words, once must conside# not only the rtratn thickness but also its distance from the seam as a 5m unit, 2m above the seam for example can be just as dcstmtive as a 30m unit, 30m above the seam.

Recent experience suggests that the possibility of periodic weighting can be indited by examination of bonhole sections. Indeed, ACIRL have tested this by making predictions of possible weighting wnditions prior to monitoring and initial results indicate the success of these predictions.

The o v 4 concept suggested is that variable weighting tends to fit somewhere khmen the light and heavy loadiing conditions of periodic weighting and can still conceivably be beset by geomechanical problems

11th I n m d Confemn~e on Ground Cmtrd in Mhing, The Unhuily of Wdlonpong, N.S.W.. July 1982.

Page 4: Australian Longwall Geomechanics - A Recent Study

Russell C. Frith. Alex M. Stewart & David Rice

Page 5: Australian Longwall Geomechanics - A Recent Study

Russell C. Frith, Alex M. Stewart & David Price

if local conditions dictate. Rgures 4 and 5 illustrate measurements taken for both variable and periodic weighting nspectively. nK fact of the matter though is that the weighting mechanism is an INEVITABLE CONSEQUENCE of mining as it is determined by the cave abiiity of the saata, not the support rating. Whilst it can be predicted it cannot be prevented, the question therefore is "can it be controlled and if so how?"

The weighting mechanism due to longwall mining and the caving of overlying strata will manifest itself at seam level as a roof convergence nte which when combined with the Jpeed of mining determined the absolute convergence level of the roof over one mining cycle. For the purposes of this discussion it will be assumed that the speed of mining is reaswably consistent form mine to mine. Hence, convergence rate can be used in the place of absolute convergence.

It is believed that the degree of face spall is driven by the convergence rate of the roof, and observations during monitoring bear out this statement. The only situation whereby roof convergence may not cause face spaU is in conditions whereby the floor material is sigdticantly weaker than the coal and the coal has little or no structural features eg Great Northern Seam. This however is the exception to the rule.

The major consideration with nspect to face a@ is its effect upon productivity andlor safety. Essentially, it is necessary to consider the block size of any face spa11 and this is determinable from examination of the structural features of the coal. A well cleated, frachlred coal will Jppll in small pi-, which pose no threat to safety and the A.F.C. can easily hansport. Convenely, a coal seam with widely spaced major cleats and only minor fracturing will tend to spall in large lumps (mares in dimension). This poses not only a threat to safety but also can d u c e productivity if the A.F.C. cannot clear the lump. Blasting may be required in extreme conditions.

The immediate roof between the tip of the canopy and the face line is the most sensitive area with

m p c t to the convergence of the roof. Due to cantileva bending &on, using the coal face (and npt the so-dkd breaking-off line at the mr of the supports) as a fulcrum, high mmprusive stress an genartcd on the undersides of the roof strata and morr reriously, high tensile stress are gmented on the uppr mhcca of the individual beds. The magnitudes of there stram an directly related to the degree of roof mvexgence. Face spall, increasing the effective tip to faceline distance, compounds this effect. The . . Wehhood of a roof fall will be directly related to the strength and structure of the immediate roof stnta. Stnu3ure of the immediate roof is of grratcr importaw than actual strength. Cantilever theory indicates that the amount of convngcnce a kam can sustain before failure is directly related to the tensile stnngth and the beam thickness. While tensile strengths of rocks may vary by a factor of up to fifty, W i g thickn- may vary from a few millimeters to tens of nwm - a much more sipniheant variation. Rooffallscanpersistupwardsforsevenl~ifthe immediate mof contains no competent shata units to abate strata failure.

As with the weighting mechanism it is ~LXI evident that the type and magnitude of face spall and p~Jibi i ty ~d =tent of roof falls can be cvalu~ted from examination of bodole d o n s and lor any exposum of therelevant immediatcstnUa and coal eithes in the form of existing undergmund workings or adjacent opn a t site. Simple e q h a h g logic and previous experience is more than adequate in making such judgments.

Won d i g in detail ACIRL's appmch to design it is worth making a few comments wnceming traditional design mcthoda. There arc

commonly used and referred m design mcthodr ranging from simple analytical models (detached block, Wilrcn [3]) to complex numaid models (finite e h a ~ t s , Pollington and Isaac [4D. Attempts have atso been made to relate mePJund khpviour to geological conditions by slatistical analysis (Pcng et at [S]). Smart [q has combincd bulking factor condddons with gwlogid d t i m to fonnulatc a 'yidding foundation model'. This model identifiu different caving mechanics Pccording to the natun of the overlying stnta.

methods in detail but simply to make geaaPl comments mmming their validity and short-comings.

1 lth h(emaW Conference on Gmuml ConW in Mihg, The Unhrersi(y d W o l b n ~ N.s.w.. ~ u t y 1982.

135

Page 6: Australian Longwall Geomechanics - A Recent Study

Russell C. Frith, Alex M. Stewart & David Rice

Taking the detached block madel in isolation, it is possibly the most used method although its limitations have been recognisecl as evidenced by attempts to supersede it. The basic assumption is shown in Figure 6. The model balances the mass of the detached block with support rating with any werlying strata load bridging from the solid d to recompacted goaf. The height of the block is defined by the bulking factor of the goaf and the height of caving required such that the void left by coal extraction is filled up.

\

UIW. I.., I.II,,.C."I

I., 1 1 1 1 1 ,(10.1,0" m,

There are several major problems with the detached block model which are now outlined:

(0 no acwunt is taken of the geology .and structure of the werlying strafa.

tendency to concentrate on the immediate roof and discount the possibility of higher caving mechanics.

(iii) based on bulking factors (which is not definitive) and variations in wnditions (i.e. non-Eumpean) are explained by different bulking factors not varying geological conditions. By the term "non-Eumpean" it is assumed to refer to

wnditions suociated with massive werburden and periodic weighting which is not always found in the cyclic mcasuns of Europe.

(iv) weruse of trial and error and imppmpriate use of fa*on of safety.

Peng refaring to the USA states that "the g e n d p d c e in this country is that if supports of lower capacity do not work out, the next one of higher capacity is used' m. Similarly, Facton of Safety between 1.5 and 2.0 are quite common in detached block designs which are high for the design of a structure which may only be required to be stable for sevual hours in g e n d .

The approach of Smart [a is partially based on bulking factors but recognises a need for adjustment to acwunt for non - European conditions. It also defines two forms of caving, namely bulking factor mmUed and parting plane wnmlled and introduce8 the design conccot of a Yiddine Foundation Modd (see Rpure - 7). l~formation as g t h e specifics of the approach are spanc and it is therefore difficult to scrutinise its application. The major perceived limitation is the fact that although account is taken of massive overlying strata the effect of any bridging beds is ignored (see %re 7) even thwgh they are shown as mveqing strata layers. Note also in Figure 7 that then is no coaceqt of strata failure either in the immediate roof or the bridging beds.

The approPch dchilcd by Peng et al[q is based on empirical guidelines attained from statistical analysis of measured behaviwr in the USA. It would appear that some of the factors of importance are rewgnised in this paper. The major drawback with their analysis is that whkt it is more complex than the dctpchcd block model it still simplities the problem to a level which makes extrapolation to other sites fraught with 'fticulty. It also ignores the phenomenon of face spall and the question of productivity qccts.

If one accepts that t h e g f f l d i s of longwall fa- an driven by the failure of overlying strata and that face @/roof falls are driven by the resultant roof convergence, then fundamental problems in terms of a deslgn model become evident:

numerical models an not totally suitable in the prediction of strata failure and it is very difficult to assign realistic values to rock mass anngths if indeed this is an acceptable practice

Mhhrg. The Un- of Wollongarg. N.S.W.. July 1982

Page 7: Australian Longwall Geomechanics - A Recent Study

(i)

Russell C. Faith. Akx M. Stewart 11 David Riee

numerical models are not sufficiently well developed to allow the accurate computation of induced displacements from the resultant stresses

measurements by ACIRL show that the longwall support problem is dynamic rather than of static equilibrium as usually assumed (see Fire 8). Numerical models whilst being able to incorporate time dependentlcmp laws, are not proven as being advanced enough to accurately solve dynamic problems.

Table 1 illustrates the applicability of the NCB Reduction Instruction (based on a detached block conapt) for the yield load of a support to measured behaviour on Australian longwalls. Whilst it is not intended to suggest that the N.C.B. PI is used for the design of Australian longwalls it will m e to illustrate the disparity of the detached block approach and the reality of support behaviour on Australian longwalk. It must be rememberrd that under the detached block concept a high factor of safety is applied and that acmrdingly the support should theoretically not go into yield. It is quite clear from Table 1 that the themtical SLJYs are gcMally 50% l o w than the actual SLD's, yet the majority of the longwall fam meanucd are either consistently at yield orpuiodically at yield. It is concluded that thae must be some fundamental m r in logic in the daachcd block approach given the wide disparity bmeen themetical and actual support behaviour.

One final comment &tes to the support specikath in terms of SLD for the four of the most meat Ausenli longwalls supplied by three different man-. All four suppan specifications are 100 Tim' or wry close to that figure. The question posed ia rhea 'if three manufactures can arrive at the same an- to four vastly different problems, why do we cvm need a d&n method for longwPU support ~ ~ " ? Thia q W o n will be answered and put into context in the nrnsinder of this paper.

As a mult, ACIRL have reuvaluated suppon dedgn rcquhmmts and are f0rmuLcing a new rpporch brvd on a sowd conoephlnl model of what hsppauuoundalongwallfaaudwhatthcsuppon h required to do to allow mining to pmaed uohinderrd. As a guide scvual fundamental a r a M o n s need to be n-itaated:

(9 at prcscnt, wardley of the approach Faken on SLD of 100 Tim2 will be arrived at. This adm out of the cumntly available technology at a raLIonable price. The cod of SLD's above this due may increase dispropoltionatdy.

@I so& of the high cmwgencc rates monitored ncently (with high suppon ratings) suggest that even with unlimited finanao, a suffidently large support couldnotbemanufauuredtonducethc

I lth lntunahd C d e m a on Dmund Contml in Minhg. The Univenity d WdlanOon0, HSW, July 1882.

137

Page 8: Australian Longwall Geomechanics - A Recent Study

RuseU C. Fritb, A b M. Stewart & Davld Price

convergence rate to a insignificant level. In other words, there are inevitable consequences involved with longwall mining in certain conditions over which we have no control. It is believed that such conditions only relate to the heavy loading phase associated with periodic weighting. With regard to variable weighting it is thought that support rating may be more of an issue as due to the lower convergence rates measured, there may be Jcope for impmving face conditions by increasing support ratings.

when investing tens of millions of dollan on capital equipment alone, there should at least be some guidelines available to establish the likely mining conditiondproblems such that the return on that investment can be assessed.

ACIRL'sapproach is to utilise empiricism based on the data bank of support /face performance data being collected from around Australia at p m t (under NWDDC sponsorship). Distinct patterns are Uaerging from this work which relate overlying geology, local geology, coal structure, working methods and other factors to geomeehanical problems and resulting productivity. This will allow judgements to be made rclacing to likely conditions in coal seams which are as yet unmined or undeveloped. The impetus bebind a fint pass design is not to specify the support rating but to provide meaningful arguments by which tbc mias planner can assess the risk of investing large amounts of capital in terms of the expected ntums.

9 illustrates the logic flow by which such a design approach would follow. At preaenl it must be messed that the database of information is based on a limited data set (14 faces) and that f W work is intending to expand it by monitoring additional Australian and overseas lmgwalls.

In addition the arguments presented for the geomechanical assessment of longwalls and their productivity, ACIRL is presently examining other fcature~ of longwall supports as part of the same m D C pmjst. Many features of longwall mining and l up port appear to be poorly understood to some degree and wanant examination. Whilst it is not within the scope of this papa to discuss them, they are

listed simply for mf-:

- amt&Ct advance -Po-* - canopy bchaviou~ during support advance - rpeod of lowa-advanamct (LAR) cycle times in nlationtoUlehaulagespeedoftheshePnr ~ v e a c t a n d i t s c o n t m l - g ~ yiCld valve opaation and canopy damage in - pedodio wdgbtillg d t i o m - ~ I e g v a u r f b l u i c g ~ ~ &kt of csaopy debris - ~ t 0 f d c J i g n ~ ~

This papa has highlighted the gwtcchnicd huea involved with longwall support niting ~peciflcation and assessing the potential foI high pmluaivity lmgwalhg. Intensive monitoring has identified distinct weighting mechanisms which are prcdictabk from the overlying geology and thcir effea is sscesrable from near-- geology and the scam rtrucRue itself. PurUKrmon, it is concluded that in some instances the available support load densities cannot adequately control the indumd roof convergmccs and that gumwhanical problems will occur from time to time. Mines m y exprima these pmblems, but under c& conditions, high productivity can still be maintained.

Page 9: Australian Longwall Geomechanics - A Recent Study

Russell C. Frith, Alex M. Stewart & David Price

The design approach will be based on empirical methods using a database which is currently beiig established. The use of numerical methods has been carefully examined, but rejected on the basis that the problem does not lend itself readily to their application. The database itself is at present relatively small, but monitoring in the near future is planned to better establish its validity.

The real emphasis of the design approach is not necessarily to define a support rating (that may be dictated by other non-geomechanical factors) but to assess the geomechanical risks to productivity with that available support technology. This will give the mine planner meaningful arguments by which to assess the capital outlay against the likely returns and make a fundamental decision as to the suitabiiltiy of the seam in question to longwall techniques. Experience around the world indicates that not all coal seams are suited to longwall mining and that some form of rationale needs to be applied to this question.

1. McCowan B. "Factors limiting production and research requirements for Australian longwalls" Jour. Mines, Metals and Fuels, June-July 1989 Vol, 37 # 617 pp 226229.

2. Whittaker B.N. "A review of prognss with longwall mine desgin and layout" Proc Conf. State-of- theArt of Gmund Control in Longwall Miniig and Mining Subsidence. SME 1982 pp 77-84.

3. Wilson A.H. "The stability of underground workings in the soft rocks of the Coal Measures" Int. Jour. Min. Eng. 1983 Vol. 1 pp 99-187

4. Follington I.L., Isaac A.K. 'Failure zone development above longwall panels' Ming Science and Technology, 1990 Vol. 10 pp 103-116.

5. Peng S.S., Zhu D.R., Jiang Y.M. 'Roof classification and determination of the support capcity for the fully m s h a n i d longwall facd Jour. Mines, Metals and Fuels, June-July 1989 Vol. 37 # 617 pp 289-296.

6. Smart B.G.D., Aziz N.I. "Influence of caving in Hint and Bulli scams on powered support rating' Fmc. Conf. Gmund Movement and Control related to Coal Mining. Aus.1.M.M. Aug 1986 pp 182-193.

7. Pang S.S. "Coal Mine Ground Control" John Wiley and Sons, New York 1987 p 243

1 lth Intemahal Conference on Grwnd Control In Mini. The University of WollonQonQ. N.S.W.. July 1892.

139