Issuu

Australian Coal Geology

CONTENTS Volume 3

Part 1

Colin Ward: 1

Introduction .

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

Phillip Harman: High resolution seismic reflection techniques in the coal industry . . . . . . . . . . • • •

David King and Stewart Greenhalgh: . . . . . . . . . . . . • •

In-seam seismic techniques for coal exploration . . . . . . . . . . . . . . . . . . . .

Mini- SOSIE seismic profiling of coal deposits in Australia .

M.W. Asten, Karel J. Grezl and L. Leung:

K.G. McCracken and P.J. Mathew: Bolt-hole and borehole-core logging instrumentation . . . . . . . . . . . • • • • • • •

R.I. Renwick: The uses and benefits of down-hole geophysical logging in coal exploration programs . . . . . .

Brian Groves and Edward Bowen : The application of geophysical borehole lugging to coal exploration . . . . . . . • • • • •

Volume 3 1981 Part 1

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 :

Ian Blayden, Anton Crouch Colin Ward

Address for editorial matters:

Address for other matters:

Anton Crouch McElroy Bryan & Associates Pty Ltd P.O. Box 387 Milsons Point N.S.W. 2061

Michelle Smyth CSIRO Fuel Geoscience Unit P.O. Box 136 North Ryde N.S.W. 2113

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 1

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

Volume 3

Part 1

Colin Ward: Introduction .

Phillip Harman: High resolution seismic reflection techniques in the coal industry .

David King and Stewart Greenhalgh: Mini- SOSIE seismic profiling of coal deposits in Australia .

M.W. Asten, Karel J. Grezl and L. Leung: In-seam seismic techniques for coal exploration .

K.G. McCracken and P.J. Mathew : Bolt-hole and borehole-core logging instrumentation .

R.I. Renwick: The uses and benefits of down-hole geophysical logging in coal exploration programs .

Brian Groves and Edward Bowen : The application of geophysical borehole logging to coal exploration .

Issued , May 1982

Introduction Colin Ward Coal Geology Group Chairman, 1979- 81. The symposium on "Progress in Coal Geology", held between 26 and 28 August, 1980 brought together some 220 registrants from industry, government and academic institutions through-out Australia, New Zealand and south-east Asia. A total of 22 papers was presented, supported by poster displays, technical exhibits and field excursions, with the object of presenting an up-to-date review of technical developments and recent experience in coalfield exploration and mine geology. The need for a symposium of this type was first suggested by the Standing Committee on Coalfield Geology of New South Wales. The Coal Geology Group, with its wide cross-section of membership, provides an ideal forum for meetings of this type and following the success of this function, arrangements are now under way for another to be held in Melbourne during 1982. This issue of the Journal, the first under the new title "Australian Coal Geology", includes the full text of six of the technical papers from the "Progress" symposium. It is no coincidence that these all deal with aspects of geophysics in coalfield investigations, since the application of geophysical methods represents one of the most significant advances that has taken place in the industry over the last few years. Three of the papers deal with seismic methods, directed as much towards mine planning as to regional exploration, and three deal with the geophysical logging of boreholes. Such techniques are now an integral part of many programs for exploration and mine development and these papers should provide valuable information on these developments for the practising coal geologist. Special thanks must be recorded at this time to Ian Blayden, who acted as convenor for the technical sessions at the symposium, and to Michelle Smyth for handling the administrative work connected with all those registrations. Dick Austen, representing, among others, the Australian Coal Association is also thanked for providing the keynote address, and so too are the chairmen of the various sessions. Lastly, but most significantly, we would like to thank the authors of these papers for the trouble they have taken with their contributions.

Hunter Valley excursion participants, Foybrook open cut, 29 August 1980

High resolution seismic reflection techniques in the coal industry Phillip Harman The Broken Hill Proprietary Company Limited, G.P.O. Box 1383, Brisbane. QLD 4001

ABSTRACT Seismic reflection techniques have been used in Australia for many years to assist in coalfield evaluation. They have generally been used to infer coal seam continuity and to locate major faults. Field practice and data .processing were originally similar to those used in early seismic exploration for petroleum. Since 1978, however, the Broken Hill Pty Co. Ltd. has applied seismic reflection techniques similar to those first used by the National Coal Board of Great Britain, with the aim of improving resolution and locating much smaller structures than had previously been identified. Achieving more confident definition of small structures involves the use of modem seismic recording equipment and digital data processing. Field practices are modified by scaling down spread dimensions to give maximum data coverage at the shallower depths of interest. Source and detector configuration are designed to maximise the generation and detection of higher frequencies in the ground and accurate static corrections are applied to compensate for variation in surface elevation and inhomogeniety of surface weathered lavers. Approximately 100 line kilometres of high resolution work has now been carried -out on closespaced grids over Cook and Leichhardt Collieries ir1 the Blackwater area, Queensland. Line spacing on average is 200m. Faults with inferred throws of as little as 4 m to 5 m have been detected at depths of 300 m. A number of known faults have been clearly identified by the method, and a clearer understanding of sedinlentary features may also be possible. The mapping of faults and steep gradients in these mines has enabled more confident selection of mineable areas and better planning of panel layout. The potential of the seismic reflection technique in the coal industry is still relatively unexploited. The role of the method in the exploration of suitable areas will become increasingly significant as the economic inlportance of coal changes and more efficient development and exploration of reserves becomes necessary.

INTRODUCTION Seismic reflection techniques have been used in coalfields in Australia since 1963. The early surveys were designed to map coal seam continuity and structure between widely spaced boreholes and to locate any major faults which could hinder underground mine development. Generally, success was achieved, however the resolution of structural detail was limited to major faults with a throw of 15 m or greater. Underground mining in the newer coalfields in Queensland indicated a high incidence of complex faults with throws much less that this. Because of the failure of these early surveys to locate such features with the degree of confidence necessary for mine planning, the seismic reflection technique was used less frequently in exploration programmes during the 1970s. Work carried out by the National Coal Board of Great Britain (Ziolkowski and Lerwill, 1977) indicated that, by using modem seismic technology and modifying field technique, it is

possible to improve greatly the resolving power of the method and locate smaller structures with a higher degree of confidence. In 1978, The Broken Hill Pty Co. Ltd carried out the first high resolution seismic surveys in Australia, mainly over the Blackwater area in Central Queensland. This work indicated the ability of the technique to locate small structures and showed its cost effectiveness in coalfield exploration. Since then further surveys have been conducted and the method has been applied as a routine tool in this area.

DEVEWPMENT OF IDGH RESOLUTION SEISMIC TECHNIQUES The seismic technique used in surveys prior to 1970 was similar to that adopted widely in petroleum exploration before the advent of modem recording instrumentation and data processing. In order to improve the resolving power of the method, finer detail, both vertical and horizontal, needs to be recorded. To achieve this higher

frequency, shorter wavelength reflections are required on the final seismic sections with closer lateral sampling so that horizontal detail may be achieved over subsurface structures. Certain aspects of the early data gathering techniques contravened these principles because of the restrictions placed on data recording by the instrumentation and the lack of data processing.

Field Technique A typical field layout used prior to 1970 is illustrated in Figure I. Data was recorded with 24 trace analogue equipment. Twelve traces were spread out either side of the shot-point at intervals of either 17m or 34m. Groups of eight or more geophones were placed in line over each interval and continuous single-fold reflection cover along the traverse was recorded by having a shot-point every 12 traces. Shots were placed below the low velocity weathered surface layer and charge sizes ranged between 0.5 kg and 2 kg of explosive. The recorder. The first change to achieve better resolution was the use of a modem floating point digital recording instrument. These instruments are capable of recording up to 48 incoming data channels at one time with a 1 millisecond sample rate. They have a better dynamic performance and can record all of the incoming signal without distortion. With analogue recorders it was necessary to compromise the data in the field by filtering or by the use of different levels of auto-

matic gain control. Some data was thereby discarded before it was recorded. The geophone group. Each trace was recorded with a group of geophones. In conventional seismic prospecting, groups are used for two principal reasons. One is to discriminate against high amplitude, horizontally travelling coherent shot noise such as ground-roll. The other is to improve the signal to random noise ratio. With each geophone planted in a different position, the response to random background noise is unlikely to be correlated between group members and will sum destructively. Improvement in resolution is gained by recording higher seismic frequencies. Unfortunately, groups of geophones have several disadvantages when this is the aim of the seismic survey. For example, a frequency of 100 Hz has a period of I O milliseconds. A relative time displacement of only 5 milliseconds between two geophones in the same group will cancel this frequency. A shift of less than this will smear the output between two geophones in a group and lower the apparent frequency, thus decreasing the signal resolving power. Such time shifts between geophones in a group occur for three principal reasons (Ziolkowski and Lerwill, 1977): (a) moveout of the reflected seismic energy across the group of geophones. This arises because of the difference in reflection path length as offset from the shot increases.

Direction of profiling Analogue recorder Last shot point

Inset

20-30 metres Base of weathering

, 17 or 34 metre;i Velocity - 2500-3200 m/ sec

300-400 metres Base of weathering ~

Geophone Group

Coal seam

Figure 1

Typical field layout used in the recording of analogue seismic reflection data prior to 1978

5 It is particularly exaggerated if the reflector is a dipping surface; (b) difference in the coupling of individual geophones with the ground and the varying mechanical properties of the surface soil can lead to lags in response ; (c) variation in surface elevation and weathered layer characteristics across a group of geophones can lead to significant differences in reflection arrival times at individual geophones in a group. An example of this effect is shown in Fig. 2. Differences of up to 10 milliseconds occur between individual geophones. Group lengths used for earlier surveys are indicated. The smearing of these static delays into groups of geophones severely changes the character of the reflection and attenuates the high frequencies. In high resolution seismic surveys the geophone group is replaced by a single high fidelity geophone. To improve the geophone coupling and minimise background noise , particularly high frequency wind noise, it is planted in a hole about about 0.5 m to 1 m deep. Geophone Group Lengths 19 70

Figure 2

Static delays on records recorded with single geophones. These delays would be smeared into geophone groups.

It is worth noting here that one of the major disadvantages of the Mini-SOSIE* shallow seismic system is that it contravenes the above principles at both the geophone group and the source. Thus overall, lower frequencies would be expected. Shot. Sharpe (1944) showed that the seismic frequencies generated by an explosive source is a function of the cube root of the charge mass. This was discussed further by Ziolkowski and Lerwill (1977) and it was concluded that smaller charges should be used in shallow coal work for the following reasons: (a) the smaller charge generates more of the high frequencies necessary to obtain the desired resolution ; (b) the depths of interest are shallower than in petroleum prospecting and smaller charges can be used because less penetration is required. For high resolution surveys, the charge size is chosen by experimentation at the beginning of the survey and is generally the smallest charge with which it is possible to detect reflections with an adequate signal to noise ratio. In the Blackwater area, the charge used is generally 50 g to 100 g of a high velocity explosive. Spread dimensions. The geophone interval is chosen to give the horizontal detail necessary to define structures adequately. In addition, the trace spacing must be sufficiently close for horizontally travelling noise to be adequately sampled. This noise may be removed later by special data processing programs which have a similar effect to using groups of geophones in field recording. A further consideration is the distance of the furthest geophone in any one spread from the shot. Because of geometric considerations, this should not usually exceed the depth of the main reflector (Ziolkowski, 1979). With shallow data this maximum offset often also determines the maximum geophone spacing that may be used. In the Blackwater area geophone intervals used to date have been either 7 m where the coal seams are at shallow depths (a maximum offset of 168 m) or 10 m in the deeper areas (a maximum offset of 240 m). Common depth point shooting and data redundancy. The other major difference between the recording of data before 1970 and the high resolution recording is the use of common depth point shooting techniques. If, instead of moving the shot point up by half the spead length after each shot, it is moved up less than half ( say two geophone stations), then each reflection point will be recorded more than once along the line from different geophones and shots (Figure 1).

Registered Trade Mark, Societe Nationale ELF Aquitaine (Production)

After correcting for geometric effects, these different records are summed to give a composite common depth point trace. The number of traces summed to give each trace in a survey determines the " fold " of cover or data redundancy. If the survey is designed so that each point is recorded six times, summing produces six-fold or 600% cover, 12 times produces 12 fold or 1200% cover etc. Common depth point shooting (C.D.P.) has the following advantages: (a) so that the data may be stacked, geometric corrections are made for the different ray path distancesrrepresented by the various traces for each C.D.P. These are characteristic only of the reflections, and therefore reflections are preferentially stacked in phase. Other events such as refractions, multiples etc. are stacked out of phase and cancel; (b) the signal to random noise ratio is improved by the square root of the number of traces summed together; (c) the coverage of the weathered layer is double the C.D.P. coverage thus the problem of static corrections may be handled statist~cally. The chance of static errors being mterpreted as faults is thus reduced. Static Corrections

The construction of a structure contour map from a seismic survey requires that all two-way times to reflectors be referred to a chosen datum plane. The correction of seismic traces to a datum plane is known as static correction. Static corrections are generally applied by assuming ( or measuring) an elevation correction velocity and calculating the time that needs to be added to or subtracted from a seismic trace so that the shot and geophones are on the datum plane. The effect of variation in surface elevation along a traverse is thus removed. In making an elevation correction it is usually assumed that the velocity between ground surface and datum level is constant and uniform along the line. The velocity through unweathered sediments varies sufficiently slowly for this assumption to be reasonable. However, the surface weathered layer is very variable in both its thickness and velocity characteristics. Since the velocity (generally 1000 m/s to 1500 m/s) is one third to one half that of the assumed elevation velocity (3000 m/s) a considerable portion of the seismic travel time is spent in this layer. For example, in the Blackwater area where the weathered layer has a mean thickness of 25 m and a mean velocity of 1200 m/s, the one-way travel time is approximately 20 milliseconds. If the shot is below the weathered layer, this delay is 10% of the total two-way time. If the shot is above the weathered layer, it is nearly 20%. In addition, the thickness of and velocity in the

weathered layer is variable and delays of up to 5 milliseconds between geophone stations are not uncommon. In typical Perrnian sediments, the effect on a reflector of a fault of 5 m throw is a difference of 3 milliseconds in two-way time. Time delays of several milliseconds due to variations in weathering cannot be tolerated and every effort must be made to reduce their effect. In the early coalfield seismic surveys statics at shotpoints were calculated by placing the shot below the weathered layer and measuring the travel time from the shot to the surface. Intermediate geophone station statics were calculated by an examination of refracted fust arrivals. Although care was taken with the static corrections they were probably a greater limitation than was realised at the time because groups of geophones were used. High resolution statics are better controlled because of the redundancy of data generated in C.D.P. shooting. Shot points occur at least every four geophone stations giving an accurate travel time through the weathered zone which can be used to make the appropriate correction. In addition , redundancy of first arrival information is double the C.D.P. redundancy so that accurate statics may be computed at intermediate geophone stations. This redundancy also allows the use of "automatic statics" computer programs during data processin~ to remove statistically any residual static corrections which may remain after manual correction.

TABLE 1 Typical Processing Sequence

1. 2.

3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

Demultiplex data and apply gain function Apply velocity filter to remove coherent noise trains such as refractions and ground roll. Sort data into common depth point order and edit bad traces Apply field statics Apply spiking deconvolution operator Apply 50 Hz to 250 Hz band pass filter Apply velocity correction for normal moveout Apply automatic residual statics Stack Apply spiking or gap deconvolution operator Apply 50 Hz to 250 Hz band pass filter Apply coherency filter to amplify selectively coherent reflections Display at selected scale

7 Data Processing

Data processing of seismic data recorded prior to 1970 was either analogue or rudimentary digital processing. It generally consisted of sorting the data, applying static corrections and geometric corrections for trace offset from the shot or normal moveout corrections (NMO), simple band-pass filtering, some amplitude equalisation and display as continuous profiles. More sophisticated data processing techniques are applied to high resolution data. Such techniques are beyond the scope of this paper but are discussed in several recent books such as that by Waters {1978). A typical processing sequence is shown in Table 1. It varies however depending on the requirements of data from different areas. The usual result is a continuous profile showing distance along the horizontal axis and two-way travel time to datum on the vertical axis.

Before 1978 Recording Instrument

Analogue 24 trace frequency modulated recorder with feedback automatic gain control.

Digital floating point recorder with I millisecond sample rate into 48 traces.

Spread Dimensions

200m or400m split spreads with continuous single fold recording. !Sm or 30m between geophone groups. 200m or400m between shot points.

Sm to !Orn geophone station spacing. Shot interval 20m to 40m to give 6 fold recording.

Geophones

Eight 30Hz geophones per trace spread over !Sm or 30m on the surface.

One I00Hz geophone per trace.

Shot

0.Skg to I kg below zone of weathering.

0.06kg to 0 .12kg below zone of weathering.

Statics

First arrjval analysis.

First arrival analysis and residual correction.

Processing

Analogue and some elementary digital.

Full modern digital processing.

Comparison

A comparison between recording techniques used in early seismic surveys and those used in a modern high resolution survey is given in Table 2. A segment of data recorded in 1970 using analogue techniques is compared in Figure 3 with high resolution data recorded in 1978 over the same structure. The period of the reflections in the high resolution data is approximately 7 milleseconds, equivalent to a dominant frequency of 130 Hz to 140 Hz. The equivalent for the 1970 data is 17 milliseconds, equivalent to 60 Hz to 70 Hz. The resulting improvement in resolution

200m Depth

High Resolution

TABLE2

Comparison of seismic survey techniques before 1978 with high resolution techniques.

·I

I·

200m

·I

100

200 I

~ -..i 300

.·-··--;-;_

0 ·2

•• ,..

....~-~!'.

·--

r~· ~~M ~r~

400

l ~~~:flt . .;,,_._,, ... ~IJ

.......:

L - ~'aJ

1970 Figure 3

• ·,

Comparison of data recorded using high resolution techniques with data recorded using earlier methods.

Structural detail available at 250m depth

Cost per line kilometre ( IOm geophone interval}

High resolution seismic

Possibly down to 3m. Some idea of geometry.

$5000

Mini- SOSI E

Possibly down to ! Orn. Some idea of geometry.

$3000 - $4000

Open hole drilling at 50m intervals - no geophysical logging

Depending on geological environment - no idea of geometry.

Continuous miner

o advance detail

0 0

0 1

$30000

(Cost of encountering unexpected structures of throw greater than seam thickness including development cost and lost production .)

0 2

sees

coal seam

TABLE 3

metres

allows the structure, which would have been interpreted as a syncline from the original seismic data, to be interpreted as a complex series of reverse faults. This difference in this interpretation may be critical to coal mine planning.

IDGH RESOLUTION SEISMIC RESULTS OVER KNOWN GEOLOGICAL STRUCTURES Since the first high resolution seismic reflection surveys were conducted in the Blackwater area in 1978, seismic data has been recorded over a number of known structures and sedimentary features. More importantly a number of structures predicted from s.eismic data have been encountered in mine workings. The fault shown in Figure 4 was encountered early in the life of a colliery at Blackwater. The important feature of this structure is that its overall width is only about 20 m and although on the face of the main fault the throw is some 7 m to 10 m, the nett displacement is very small. At the time it was encountered in the mine, the geometry was unknown and months were spent cutting down into the graben only to find the coal seam at the original level on the other side of the structure. If the geometry had been known at the time, development could have proceeded at the one level across the fault. The geometry of the fault is clearly defined by the seismic data, and the extension of this feature away from the mine workings has now been accurately mapped by seismic work as well. A similar but smaller fault is shown in Figure 5 along with the equivalent seismic record. The width of this structu.re is only 10 m and the throw 3 m to 5 m. It is however, clearly visible on the seismic section.

Figure 4

Seismic data over a 5m to JOm fault known in colliery workings.

::,.

[. 0

metres 0

Figure5

10 U

...,,..

Seismic data over a 3m fault known in colliery workings.

9 Depth 0

100 m

Time

0. 0 ----,-rrfirn,,.....-rrTTirrrTTTT"rrfin-TTTTT_,rrMTTT"rrfin-TTTTTTTT",.......TTTrrfin-TTTTTrTTirrrTTl"Prrfin-TTTTT"rTTirTTTTM

100 0.1 200

300 Metres

Figure 6

Depth

Seismic data over a 20m fault predicted before being encountered in mining.

-~~-

Tome 00

100 01 200

300

Figu;·e 7

Major colliery boundary fault 1 00m.

A much larger structure of similar type is indicated by the seismic section in Figure 6. The throw is in excess of 20 m with very steep grades on the left side of the fault. The overall width of the structure is less than 100 m. Management was warned about the existence of this structure following the seismic survey in the area. The mine workings approached the fault from the right-hand side of the section, and drilling ahead from the workings at the fault plane confirmed the geometry indicated by the seismic work. A comparison between mine levels in the workings and coal seam depths determined from a bore at shot point 943 on

the section failed to reveal any evidence of this major, panel-stopping structure. A seismic section across a major fault at the boundary in one of the collieries is shown in Figure 7. This fault, with a throw in excess of 100 m, was known from drilling prior to the establishment of the colliery, and the seismic line was placed to confirm its position and determine its geometry. It is clearly not a simple normal fault, and movement during sedimentation may be inferred by the draping of beds across the fault plane .. The orientation of this feature was originally inferred from drilling and photolineament mapping. The seismic survey

10 confirmed the fault's existence and showed that, rather than being simple it is complex and part of an en-echelon system along strike. Based on seismic results, the orientation of the fault was re-interpreted, considerably increasing the accessible reserves of the colliery. A reverse fault of 10 m throw is interpreted from the seismic data in Figure 8. A close examination reveals a small event which is probably a reflection from the fault plane. This fault was undetected in the mine workings until a drill hole from the floor intersected the same coal seam 4 m below. The interpreted structure is depicted below the seismic record.

Figure 9

Interseam interval widening suggesting a channel sandstone between coal seams.

LIMITATIO S OF THE TECHNIQUE The above examples have shown the effectiveness of modern seismic techniques in locating potential mining hazards such as faulting. This has been achieved by adopting a field approach aimed at generating and recording seismic waves with higher frequencies and shorter wavelengths than those usually used in seismic prospecting. Despite this demonstrated effectiveness, the limitations of the technique must be remembered when a seismic survey is being contemplated. Realistic Limits of Detection ----.....:;

Figure 8

Predicted reverse fault confirmed by drilling in the mine floor.

Sedimentary Features

A change in the interval between two seams is implied by the divergence of the upper two reflectors in Figure 9. Drilling information has confirmed an increase in interseam thickness at this location along with an increase in sand/shale ratios. Roof problems were experienced in the mine in this area and similiar areas of potential hazard have been located elsewhere from seismic data. An example of seismic data over seam splitting is shown in Figure 10. The area of splitting is marked by a change in the character of the seismic event from one side to the other. When the separation of two seams is sufficiently great, reflections from each may be distinguished.

Even at seismic frequencies of 150 Hz, for a velocity of 3000 m/s, the seismic wavelength is 20 m. The top and bottom of a 3 m coal seam cannot be resolved as separate reflections, because the reflections from the roof and floor interfere with one another. In addition, if two seams are close together then their reflections will interfere with one another. The detectable limit of a structure such as a fault depends on several factors including: (a) data quality, as defined by the signal to noise ratio and the frequency content of the reflection on the final processed seismic record ; (b) the geometry of the structure and the orientation of the traverse line in relation to it. Sharply terminated structures such as faults will probably have a clearer expression on a seismic section than a broader feature such as a roll. In addition, structures are most clearly defined by lines which cross them perpendicular to the strike. In early seismic surveys on a prospect, the strike of structures is often unknown and orientation of lines may be unfavourable for effective definition;

11 areas where the surface is covered by thick, loosely consolidated material such as surface sand, high frequencies are attenuated in this layer, reducing the resolving power of the technique. Particularly hard surface cover such as basalt also prevents the use of surface seismic techniques because the energy is trapped in these layers and does not penetrate to the underlying coal measures. In other areas, limited surface access, different drilling conditions etc. can reduce the resolving power and therefore the effectiveness of the method.

Figure 10

(c)

(d)

Seismic reflection character over an area of coal seam splitting.

depth of the coal seam. Higher seismic frequencies are attenuated at a greater rate with distance travelled than lower frequencies (Farr, 1976). Thus as the depth of a coal seam increases there is a corresponding lowering of the frequencies in the reflected energy, and a loss of resolving power; the general physical characteristics of seismic waves. Wavelength is not only proportional to frequency but also to velocity for a given frequency. The higher the velocity, the longer the wavelength and the higher the limit of detectability for that frequency.

Multiple Seam Environments Coal seams have low velocity characteristics and density. There is a very high acoustic contrast between them and the surrounding sediments, resulting in high reflection co-efficients. A high proportion of incident seismic energy is reflected, not only back to the surface but also as reverberations within and between coal seams. Such reverberation creates strong multiple events and leads to the severe attenuation of all seismic energy, particularly the high frequency components. Work by Ruter and Schepers (1978) demonstrated that below the top four or five seams in a coal measure sequence, events on seismic reflection records are no longer primary reflections but multiples and mode converted waves. These cannot be tied to individual coal seams and thus in a multiple seam situation it is only possible to map structure on the upper seams. The attenuation of the higher frequency components by the the upper-most coal seams also leads to a concomitant increase in limits of detection on deeper seams.

Surface Conditions and Geology The greatest contributing factor to poor data quality is unfavourable surface conditions. In

Dip The theory of seismic reflection assumes essentially flat lying reflectors. The techniques of field data gathering and of data processing preclude steeply dipping reflectors and thus should not be applied over areas where dips exceed 30 degrees.

COST EFFECTIVENESS The cost effectiveness of the technique will vary from one area to another. It must be weighed against the cost of either using other techniques or bearing the risk of not knowing if any colliery stopping structures exist. The cost of dynamite seismic surveys depends on many factors, the greatest variable being the cost of shot-hole drilling. Since a typical program consisting of 50 line kilometres with a 10 m geophone spacing would typically involve 20 000 m to 30 000 m of shot hole drilling, variation in drilling rates can greatly affect the overall cost. In addition, as the depth of the coal seam decreases, the cost per line kilometre of seismic surveys increases because of the closer geophone spacing and shorter spread lengths necessary to obtain data from the shallower reflecting horizons. For a typical seismic program in the Blackwater area of Queensland, seismic survey costs may be broken up as follows: Survey and line preparation 12% Drilling 35% Seismic crew 25% Data processing 15% Other (supervision etc.) 13%

100% Depending on the geophone and shot spacing necessary, costs may vary between $5 000 and $10 000 per line kilometre. These figures would however vary dramatically with changes in the cost of shot hole drilling. If shot holes are predrilled and loaded, a typical dynamite crew could shoot 150 to 200 records per day. An alternative shallow seismic reflection method is the Mini-SOSIE technique described elsewhere in this volume. A comparison between a line recorded using this technique and one recorded

12 plan before and after a seismic survey. The location of structures previously unknown in the workings has been confidently predicted. In addition areas of steep grade and previously unknown faults have been mapped with sufficient accuracy to enable detailed mine planning and development.

using dynamite high resolution methods is shown in Figure 11. Because the energy source of the Mini-SOSIE technique is applied to the ground surface a number of compromises are made which limit the recording of high frequencies. The difference is clearly illustrated in Figure 11 where the overall frequency content of the Mini-SOSIE section is lower than that of the dynamite section. The cost of a Mini-SOSIE survey is probably slightly more than that of a dynamite survey if the cost of shot-hole drilling is excluded. Clearly in areas where shot hole drilling is prohibitively expensive the MiniSOSIE technique offers a viable though less detailed seismic alternative. In areas such as Blackwater in Queensland, the increased detail obtained from dynamite surveys justifies the higher cost. A comparison of estimated costs of various currently available exploration techniques is presented in Table 3.

CONCLUSIONS This paper has shown that, by suitably modifying seismic reflection techniques, it is possible to obtain data in sufficient detail for the detection of structures which are a hindrance to coal mine and panel planning. It is possible to locate with confidence faults that had previously remained undetected using early seismic techniques and pattern drilling. The degree of success with the techique is limited mainly by surface conditions, sub-surface geology and the types of problem that need to be solved. The detection limits of the method are as yet unknown, but in favourable areas, a reasonable limit to aim for would be the detection of structures with a relief of 3 m or greater. The overall cost of high resolution seismic surveys depends heavily on the cost of shothole drilling. The cost effectiveness of the method can only be

EFFECTNENESS IN MINE PLANNING Experience with the high resolution seismic technique in the Blackwater area has shown the effectiveness of the method as a mine planning tool. Figure 12 illustrates a portion of a colliery

HiQh

Resolutlon

Mi ni

Figure 11

Dynam ite

Soll•

Sect ion

Secti on

Comparison of Mini-SOSIE data and high resolution data along the same line.

I I

I I I

' ,. I'~ • I

I I I I I

•

I I

~•

\ i

·c I

•:;

( ·~·

"--

f 'II

1• /

',, '!.. II I I 7I

/ /

"' 0

/ I/

l"IIIM I

Str11ct11r•

IEFORE

S • l ■ mlc

,. -

S11n1y

LEGEND

·-f--

- , l c -1 .. 11

-+ - -c-lUl t

-Figure 12

Structural information from a proposed mine panel area before and after a seismic survey.

judged according to the problems to be solved and the cost of alternative exploration techniques, including mining. The development of seismic technology is ongoing and future advances are likely in the areas of both data recording and data processing. More confident predictions will also result from the increased experience of coalfield geophysicists and geologists in the interpretation of seismic data.

ACKNOWLEDGEMENT This paper is published with the permission of the Broken Hill Proprietary Company Limited. The support of many people who have assisted in the development and understanding of ideas in this paper is gratefully acknowledged.

REFERENCES Farr, J.B., 1976. How high is high resolution?, presented at the 46th annual meeting of the Society of Exploration Geophysicists, Houston, Texas Farr, J.B., 1977. High resolution seismic methods to improve stratigraphic exploration, Oil & Gas J., November 1977 Harman, P. & Rutter, H., 1979. Seismic reflection techniques for coal exploration, B.H.P. Technical Bulletin, 23, 2,. November 1979 Ruter, H. & Schepers, R., 1978. Investigation of the seismic response of cyclically layered Carboniferous ror~ hy means of synthetic seismograrns, Gcupnysical Prospecting, 26,p 29

Sharpe, J.A., 1944. The effect of charge size on reflection records, Geophysics, 9, p 131 142 Waters, K.H., 1978. Reflection seismology John Wiley & Sons, Inc. Ziolkowski, A., 1979. Seismic profiling for coal on land; in Developments in geophysical exploration methods -1 (A. Fitch, ed.) p 271 - 306, Applied Science Publishers Ltd., London Ziolkowski, A. & Lerwill, W.E., 1977. A simple approach to high resolution seismic profiling for coal, presented at the 39th meeting of the European Association of Exploration Geophysicists, Zagreb.

Mini - SOSIE seismic profiling of coal deposits in Australia David King 1 and Stewart Greenhalgh 2

Earth Resources Foundation, University of Sydney, N.S. W. 2006

ABSTRACT Since its introduction into Australia late in 1978, the Mini-SOSIE* technique has found application for structural mapping in a number of Australian coal deposits. The technique has proved capable of providing useful structural information in all coal environments encountered. Data quality is most sensitive to the variability of near surface low velocity layering, but static correction procedures are available which mitigate the problem and enable routine production of medium resolution processed sections. The style of data collection and processing are such that the Mini-SOSIE technique is more cost-effective, relative to standard seismic techniques, the less the total vertical and lateral subsurface coverage required.

INTRODUCTION In the two years since its introduction into Australia, the Mini-SOSIE technique has found application in a variety of coal and mineral environments. It is thus timely to assess the scope of the technique in the exploration of coal deposits in Australia. Such an assessment is concerned primarily with the adequacy of the technique to provide structural information more economically and/or more acceptably (e.g. on environmental grounds) than alternative seismic techniques. A small part of the data collected in Australia using the Mini-SOSIE technique is presented as a guide to the scope and viability of the technique. METHOD The Mini- SOSIE technique is based on the generation and transmission of a long duration, pseudorandom series of low energy impulses using a common civil engineering earth tamper. That part of the input energy which returns, by way of refraction, reflection or diffraction, to a conventional detector (geophone) array on the surface is manipulated on recording so as to appear as if it emanated from a single impact. As many as 2000 individual impacts may be required to achieve an adequate signal to noise ratio. This paper is concerned with the reflection version of the Mini-SOSIE method. The reader is referred to King (1979) for a full description of the fundamentals of the technique. It is pertinent, however, to clarify a number of points concerning the basis and methodology of the Mini-SOSIE method : (a) the necessity for randomness in the input sequence is a consequence only of the collapse of the long-duration input sequence

by correlation in real-time. Non-real-time correlation would successfully collapse a non-random input sequence but entails, inter alia., excessive data recording capabilities; (b) the success of the real-time correlation operation depends on the fidelity with which the signals issued by the source sensor ( on the base plate of the earth tamper) can control the real-time correlator. It is necessary to monitor carefully the source sensor signals to ensure that signal shape and amplitude fluctuations do not introduce arbitrary delays into the processor control signals (see Figures 2b and 2c in King, 1979); (c) the use of multiple earth tampers does not improve imput randomness if the individual rarnmers are all driven in a truly random fashion. However, experience suggests that randomness is not easily achieved for a single rarnmer, and the use of multiple rammers does improve the effective randomness of the impact sequence generated; (d) the Mini-SOSIE technique is, of course, subject to the same geometrical constraints as any other shallow seismic technique. In parti_cular, the maximum geophone offset used should not be significantly greater than

Registered Trade Mark, Societe Nationale Elf Aquitaine (Production) Present Address: 1 Offshore Oil NL, 16 7 Phillip St., Sydney, 2000 2 Flinders University, Bedford Park, S.A. 5042

(e)

the minimum target depth at which fullfold CDP coverage is desired (see Ziolkowski, 1979). Furthermore, with regard to the maximum allowable geophone group spacing, the sampling theorem is inviolate; the impulses generated by earth tampers are band limited, at times severely so. Insofar as the ultimate resolution attainable is critically dependent on band width, frequency filtering at the detector and/or recording levels must be imposed as sparingly as possible.

EXAMPLES A number of successful Mini-SOSIE reflection profiles from the Gloucester Basin ofN.S.W. have been published elsewhere (King, 1979). However, the most impressive example available to us (by courtesy of Theiss Bros) is from the Callide Basin in Queensland. The profile shown as Figure 1 f - 10 0 -

clearly delineates the coal disposition and abrupt termination; moreover there can be little doubt that the upper and lower surfaces of the coal (see geological log, inset) are fully resolved by reflection pulses of opposite polarity. We are aware of other high quality examples of MiniSOSIE profiles which will be published elsewhere (Harman, pers. comm.). At the other end of the quality spectrum, Figure 2 shows a problematic section (by courtesy of Department of Main Roads) from the southern Sydney Basin. It is difficult to make structural interpretations with any confidence from the data shown. We remark, however, that the thick section of Hawkesbury Sandstone penetrated at the site of Figure 2 has proved equally problematic for "conventional" seismic techniques using explosive sources. Low signal levels when shooting in sandstone have generally been ascribed to a coupling problem. The data illustrated in Figure 2 provides an in-

--j

r---~~- -------- -------- ------- ------- --.--.---- ------- ---,0

Figure I

Mini-SOSJE section from Callide Basin, Queensland, with geological log inset

Figure 2

Mini- SOSJE section from southern Sydney Basin

17 teresting example of the sensitivity of MiniSOSIE data, or indeed any seismic data, to the digital processing used for production of the section. A portion of the data shown in Figure 2 is reproduced in Figure 3 alongside a section produced by a different processing contractor. The comparison serves to illustrate the well-known trade-off between bandwidth and resolution; it is tolerance of "ringing" in time (vertically) and space (horizontally) which dictates the preferred style of section. A further interesting processing comparison is shown in Figure 4 for data from the Gloucester Basin ofN.S.W. (Note that the plot polarity is different in the two sections). The thrust fault evident on the upper section is hardly visible on the lower section; on the other hand, the principal reflector in the lower section corresponds well to the main coal bed but is unresolved in the upper section. The comparison serves to illustrate the compromises involved in the processing of shallow data. The muting of refractions necessary to bring out the main coal reflector can also eliminate useful shallow structure. The necessity for tailoring processing to serve interpretational needs requires no further emphasis.

Figure 3

;';.. ...

,1/1•n

......... ., --

...

0 ·1

Figure 4

Comparison of two different processing sequences for data from Gloucester Basin, N.S. W. Time and distance scales are the same on both sections, but plot polarities are opposite

18 We remark that all sections shown have been produced with no static corrections for the effect of weathering. An analysis of first arrival refractions in these areas showed little or no variation due to weathering. This is not the case in all coalfields; where weathering variations are significant, it becomes crucial to determine appropriate static correlations prior to processing (as is the case for any seismic method).

COMPARISON WITH EXPLOSNE SEISMIC TECHNIQUES There is little value in comparing the Mini- SOSIE and explosive seismic techniques unless the comparison involves data collected and processed in some optimum fashion for the same objective (within the range of penetration of the former). Unfortunately this condition is clearly not met for some of the comparison data circulated by the Mini- SOSIE manufacturers, and reproduced by McQuillin et al. (I 979, p 20). The dynamite section shown in the comparison involves an array of shallow (1 meter) shots, which is inimical to resolution on several counts. Moreover, the group spacing for the dynamite survey is over twice that of the Mini-SOSIE survey, so that the percentage subsurface coverage is less for much of the dynamite section than for the Mini- SOSIE section. Such invalid comparisons are mischievious; there are surely those who suspect that the two sections published by McQuillin et al (1979) are simply labelled in reverse order! It is to be expected that dynamite data collected and processed in an optimum way would always be superior to Mini-SOSIE ( or any other surface source) data. A valid comparison of the two techniques conducted at Blackwater by the B.H.P. Co., and to be published elsewhere (Harman, in preparation), supports this view. However, such dogma cannot be extended to comparisons in terms of benefit - cost, especially where operating conditions are constrained by economic and/ or environmental considerations. Comparisons of benefit - cost are not easily generalized, but the advantages of the Mini-SOSIE crew in this regard are the lower capitalization, lower manpower requirements, and lower mobilization and operating costs. From an operational point of view, the disadvantages are the relatively slow production, both because of the extended ramming times, and the fact that recording systems have been restricted to 24 traces. Of course the production rate superiority of dynamite crews depends on the availability of adequate shot-hole drilling capacity, with its associated substantial mobilization and operating costs. Detailed costings of the many facets of a seismic survey demonstrate that a Mini-SOSIE operation has a clear cost advantage for short surveys such as are commonly required for coal exploration. It is beyond the scope of the present paper to detail the costs which determine critical or optimum survey lengths for the alternative techniques.

RECENT DEVEWPMENTS There are two recent developments which bear on the scope and viability of the Mini-SOSIE technique. Firstly, the availability of a sign-bit summing option for Mini-SOSIE recording offers the technique a technological advantage for data collection in areas where random noise is a problem. The signbit summing technique entails truncation of data samples to a single bit (the sign-bit), from which true amplitudes can be recovered by vertical stacking providing that the signal to noise ratios of the individual elements of the stack never exceed one (O'Brien et al, 1979); the condition is clearly met for the Mini-SOSIE technique. The use of sign-bit recording greatly extends the capacity for vertical sta<.;king, with a consequent improvement in the discrimination against random noise. A further advantage of sign-bit summing is that it opens up the possibility for parallel recording on a greater number of channels, which is desirable for 3D surveys or high multiplicity CDP coverage. The second recent development in Mini-SOSIE technology is the availability (from the Input/ Output Corporation, Houston) of a multisource processor (MSP), which enables amongst other things, Mini-SOSIE recording of up to 120 channels with standard seismic recording equipment. This development will, of course, greatly extend the production capacity (and the flexibility of operation) ofMini-SOSIE crews with the obvious associated cost benefits. CONCLUSIONS The role and viability of the Mini-SOSIE technique in the exploration of Australian coal deposits is already well established. Recent technological developments will encourage the extension of this role into the realm of longer, higher production surveys. ACKNOWLEDGEMENTS The authors gratefully acknowledges the help and generosity of Lance Grimstone (Theiss), Lubo Acimovic (DMR), Zenon Kociumbas (BMI) and Phil Harman (BHP). The research reported in this paper is supported under grants from the Australian Research Grants Committee and the National Energy Research Development and Demonstration Council.

REFERENCES King, D.W., 1979. Mini-SOSI E seismic profiling for coal in the Gloucester Basin of N.S.W., Bull Aust. Soc. Exp/or. Geophys., 10,

p 156 -163

McQuillin, R., Bacon, M. & Barclay, W., 1979. An introduction to seismic interpretation, Graham and Trotman Ltd., London O'Brien, J.T., Kamp, W:P. & Hoover, G.M., 1979. Theory of amplitude recovery from signbit recorded data with applications to seismic problems - 1, paper presented to 43rd annual meeting of Society of Exploration Geophysicists, New Orleans Ziolkowski, A., 1979. Seismic profiling for coal on land; in Developments in geophysical exploration methods - 1, (A. Fitch,ed.) Applied Science Publishers, London

In - seam seismic techniques for coal exploration M.W. Asten 1 , Karel J. Grezl 2 and L. Leung 2 Exploration Dept., The B.H.P. Co. Ltd., 1. P.O. Box 559, CAMBER WELL, VIC 3124 Central Research Laboratories, The B.H.P. Co. Ltd., 2. P.O. Box 188, WALLSEND, N.S.W. 2287 ABSTRACT A review of in-seam seismic techniques for coal exploration shows that in-seam seismic is

rapidly becoming an operational tool to assist in the development of underground coal mines. At the forefront in this work at present is the National Coal Board in the United Kingdom who have assembled equipment and expertise to facilitate regular staging of inseam seismic trials and who have claimed some success in detecting faults up to 500 m away from development headings in seams of 1 - 2 m in thickness. Some Government support has been made available to facilitate the development of inseam seismic techniques in Australia. Preliminary field trials at Leichhardt .Colliery in Queensland have given encouraging results and are the first in an on-going series of trials aimed at developing an in-seam seismic exploration system for Australia. Explosives were used to generate channelled waves within the coal seam and these were detected with high fidelity \it distances of up to 340 m. Propagation velocities and waveform attributes were characteristic of channel waves reported by other workers and frequency content was significantly better due to the use of broadband piezoelectric detectors.

INTRODUCTION The in-seam seismic technique (I.S.S.) is a geophysical measurement of sound-wave transmission and reflection which is performed at faces or along headings within a coal mine. The object of the measurement is to detect faults or other seam discontinuities ahead of the face and thus reduce production losses and safety problems which may be associated with such discontinuities. The wave types and procedures required have been the subject of modelling and practical experiments in Czechoslovakia, West Germany, U. K. and U.S.A. since the concept was first discussed by Krey (1963). However it is only in the last few years, since the development of digital seismic recorders suitable for use within coal mines, that the I.S.S. technique has moved from the realm of academic experiment to that of mine-planning tool. This paper has three aims: (a) to identify the basic principles of geometry and physics involved in the technique that influence the design of a survey and interpretation of results; (b) to illustrate an application of the method with some preliminary results from the first I.S.S. survey conducted in Australia; (c) to show with three case histories from West Germany and U.K. how the technique is

capable of yielding cost-effective predictions.

GEOMETRY OF AN IN-SEAM SEISMIC SURVEY The I.S.S. technique uses seismic waves travelling laterally within the coal seam and adjacent rock. A survey may be designed to detect discontinuities by transmission (or lack of it) through a fault or by reflection or diffraction from a fault or local discontinuity. Figure 1 shows an idealized hypothetical test area where a fault F 1 is intersected in one heading (B) of a prospective longwall development. A second fault F 2 is inferred from reduced levels of the seam in heading A and borehole H. Three possible experiments may be designed to gain data on the extent of faulting: (a) seam wave transmission from shot point S to detectors in A heading may indicate extent of the fault by existence of a "shadow zone" towards the eastern end of A heading; (b) detectors in B heading may receive seam waves from shot-point S reflected from the fault; (c) continuity of the coal seam between A heading and borehole H"may be tested by seam wave transmission ( orlack of it)

,-----------------:::---------:--------,--' A

•

/ ,f

~ /> •

•

')..,,

• 0 Km.

Figure 1

Idealized site for in-seam seismic transmission, reflection and hole-to-face surveys.

between a shot point at A and a down-hole detector set within the coal seam in borehole H. Current practice in the U.K. and West Germany is to use face-to-face or heading-to-heading transmission tests for calibration studies to determine characteristic frequencies and velocities of seam waves for the coal seam of interest, and then to use the face-to-face reflection technique to delineate faults. In-seam seismic techniques with boreholes may use face-to-hole, hole-to-face or hole-to-hole geometries. However borehole surveys performed in West Germany (Krey, 1976), U.K. (Buchanan, pers. comm., 1980), U.S.A. (Hasbrouck and Hadsell, 1976; Regueiro, 1980) and Australia (The Broken Hill Pty Co. Ltd, unpublished) are experimental rather than routine surveys.

PHYSICS OF THE SEAM WAVE Coal is a soft solid with a seismically low velocity, generally underlain and overlain by harder highvelocity rock. There is a consequent tendency for seismic energy to be trapped within the coal seam giving rise to wave-guide or resonant-cavity characteristics when seismic waves are generated within the seam. The resulting waves are termed seam waves ox channel waves and have quite different physical properties to the better known compressional and shear waves which propagate in an unbound solid and are used in all surface seismic exploration. Theoretical and model studies on the physical

properties of seam waves have been published by Dresen and Freystatter (1976), Freystatter and Dresen (1977, 1978), Guu (1975), Su (1976) and Peterson (1979). It is sufficient here to note five general physical properties of seam waves: (a) the waves propagate in narrow frequency bands, predictable from resonance properties of the seam "cavity"; (b) multiple modes of wave propagation, each with a different velocity, are possible; (c) for each wave mode, wave velocity varies with frequency, resulting in wave dispersion; ( d) two modes of wave motion are possible, Rayleigh waves where particle motion is elliptical in a vertical plane parallel to the direction of propagation, and Love waves where particle motion is tranverse to the direction of propagation in the plane of the coal seam. These two modes are illustrated in Figure 2; (e) the disturbance associated with a seam wave is guided by the seam, but not entirely contained within it. Thus a fault with a throw equivalent to the seam thickness will break the coal seam entirely but will only partially block a seam wave. This complicates the interpretation of transmission data. As a general rule the higher the frequency of the seam wave, the more completely is the seismic energy confined within the seam. The first three characteristics are evident when surface seismic records and in-seam records are compared. Figure 3 shows a set of unprocessed

TRANSV ERSE

VERTICAL

RA DIAL

80 100 120 140 160 180 200 msec

Figure 2

Difference in particle motion between the Rayleigh and Love classes of seam waves (reproduced from Krey, 1976).

· 50 0

j ' 250

500

750

IOOO

1250

FREQUEN CY

I Hz l FIELD MONITOR-SURFACE SEISM IC

100

200 msec

300

/v I~

---

Figure 4

r'I /

7 h ,__

IQ()

250

500

FREQUENCY ( Hz )

Figure 3

Records and a representative spectrum from a surface seismic reflection survey.

seismic records produced on a field monitor during a surface seismic reflection survey. Wavelets visible show distinct edges and ·are one or two cycles long. Figure 4 shows a set of in-seam

Records and spectrum from an inseam seismic survey at Leichhardt Colliery. Shot-detector distance is 126 m.

seismic records recorded in Leichhardt Colliery, Queensland. No reflections are believed to exist on the seam-wave records; the first event at 33 milliseconds is a compressional head-wave transmitted through roof and floor rock, while at least four later events at 84 milliseconds, 96 milliseconds, 123 milliseconds and 140 milliseconds represent different modes of seam-wave propagation. Unlike the unconfined wavelets shown in Figure 3, velocity dispersion of seam waves causes the wavelet associated with each mode to be spread out in time to a length (in this case) of 5 to 10 cycles. The physical characteristics of seam waves affect all stages of an in-seam seismic survey. To identify the different wave modes requires the use of two or three component detectors. This complicates the logistics of the survey and decreases the rate of coverage. Useful frequencies of seam waves extend up to at least 600 Hz and possibly to 1200 Hz, i.e. beyond the capability of existing conventional seismic equipment. The particular frequency-dependent velocity and dispersion characteristics of seam-wave wavelets require the development of special computer processing. display and interpretation techniques which are currently unavailable in commercial seismic processing centres.

IN-SEAM SEISMIC SURVEY: LEICHHARDT COLLIERY Published data on seam wave characteristics in U.K. and West Gennany relates to seams ranging

24 Accererometers

JUNCTION ,...___ _ BOX

F. M. TAPE SAFETY BARRIER f-----~ RECORDER

FILTERS

BOX

Blaster CHART RECORDS

CHART RECORDS

Figure 5

Instrumentation configuration as used at Leichhardt Colliery.

from 1 m to 3 m in thickness. Since seam thickness at Leichhardt Colliery ranges from 6 m to 7 m, the survey was designed to identify wave characteristics in the different geological conditions as well as to acquire transmission and reflection data. The instrum~ntation used was set up as shown in Figure 5, being an adaptation of a system developed for monitoring of microseismic noise in the mine (Grezl, Leung and Ahmed, 1980). Threecomponent piezo-electric accelerometers and preamplifiers were used as detectors, assembled in lengths of P.V.C. tube set and grouted into the coal face. The frequency response of the accelerometer and preamplifier is flat from 1 Hz to 5 kHz. The signals from the preamplifiers and a shot-break detector geophone were transmitted to mine surface on a multicore cable, passed through a safety barrier box and recorded on an 8 channel FM tape recorder having a flat ·frequency response from DC to 5 kHz. A digital oscilloscope was used for monitoring, recording and replaying of data. Two approaches to data processing and display have been used; first by replay through variable analogue filters to the digital oscilloscope, a spectral analyser and a chart recorder, and second by replay into a Geometrics ES1210F digital portable seismograph with digital tape recorder coupled to a desk-top computer. Thus digital filtering, enhancement and graphics techniques were available. An advantage of the instrument configuration was the minimization of the quantity of equipment needing modification and certification for underground usage. A disadvantage was the reduced mobility due to routing requirements for the main transmission cable, particularly in extensive underground workings. Figure 6 shows two survey configurations used

COLLIERY

LEICHHARDT

t • •hot point • occel e rometer 111• -!-fault 0

100

200

300

metru

Figure 6

Plan of Leichhardt Colliery and LS.S. survey test sites.

in Leichhardt Colliery. A pair of triaxial accelerometers were installed in mid-seam and quarterseam positions in a southeast and a north heading. One hundred gram charges of Morcol (low velocity explosive) were detonated at the marked shot-points in holes 2 m deep, drilled mid-seam. The shot-point geometry was designed to give transmission data south of heading M and west of heading T, and possible reflection data between heading P and fault F. The fault is intersected in workings in the east panel of the mine,

25 and is evident on surface seismic reflection records 1 km east of the mine. Its extent and orientation west of the east panel are unknown. Prior to shot-firing, experiments were conducted using a sledge-hammer source on a steel bar placed in shot holes. This source generated wave modes in a frequency band extending one or two octaves above that generated by explosive charges, and dispersed wavelets at frequencies up to 1200 Hz could be observed over transmission distances up to 50 m. The previous maximum frequency re-

msec 0

400

800

\ •,. - ~ -~h ' R~~~wk! ~f~~YJE~~t~~±::5·~,~ti:t:rtj ~:_,~. -~.i +±C . ~ ¥ - < --W P=la - 'f-4.!' - .. ••

~f~; ~

ported in in-seam experiments is 600 Hz, this being close to the maximum recordable on most present day seismic geophone-recorder combinations. The useful range of the hammer source was 120 m for transmission experiments. This is too short to be useful for reflection surveys as presently envisaged, but it is conceivable that the high frequencies generated may in future be useful for short-range high resolution studies of seam discontinuities. Signals generated by explosive charges were clearly received over transmission distances up to 340 m (the maximum available). The spectrum of received signals extends from 50 Hz to 500 Hz with a strong component in the band 90 Hz to 110 Hz. Figure 8 shows a set of records from shots in headings P and Q which have been filtered and amplified to enhance this frequency band. At least three modes of direct arrivals, with group velocities of approximately 1.4 km/sec, 0.95 km/sec and 0.65 km/sec may be identified. The line Don Figure 8 joins arrivals of the 0.95 km/sec mode.

II.

cool HOm

·-a

100 o,

>110 - uo,. 00

120 .,

140

160

166

b Figure 8

176

Figure 7

LS.S. records from shots sited in headings P and Q, Leichhardt Colliery. The component displayed is horizontal, normal to rib, bandpass filtered to 80 Hz to 125 Hz.

( a) (b)

LS.S. test site, West Germany transmission records; shotpoint 2 to geophone spread B ( c) transmission through fault; shot-point 3 to geophone spread A (d) reflection records; shotpoints between 2 and 3, geophone spread B (reproduced from Ruter and Schepers, 19 79)

26 Given the observed velocity of direct arrivals on Figure 7 and the hypothesized faul t position shown on Figure 6, it is possible to calculate the time at which reflected wavelets should appear on the seismic records. These calculated times for the 0.95 km/sec mode are shown on Figure 7 as the line R. Some .correlation with low amplitude wavelets on the records is evident. This analysis is preliminary and cannot be regarded as conclusive evidence for the detection of reflected seam waves or for the existence of a fault at its marked location. Further processing of all components of the data will allow firmer conclusions to be drawn.

IN-SEAM SEISMIC TECHNOLOGY IN WEST GERMANY I.S.S. surveys using the face-to-face reflection configuration are currently in a three-year evaluation phase in West Germany (Brentrup, 1979 ; Barth, 1979). A DFS-V digital seismograph has been modified to flame-proof specifications, and two component geophones have been adopted to assist in wave-mode separation (Ruter and Schepers, 1979; Millahn and Arnetzl, 1979). Figure 8 shows a case history from Ruter and Schepers (1979) where the influence of a known fault of throw 3 m to 4 m in a seam of thickness 2 m to 2.5 m was examined. Figure 8(b) shows transmission records for waves travelling from shot point 2 to geophone spread B. The first arrivals are from compressional waves travelling in rock but the second arrivals (travel time 0.1 sec to 0.2 sec.) are dispersed seam waves. Transmission waves recorded at geophone spread A from shot-point 3 are shown in Figure 8(c) ; the first arrival compressional waves exist as before, but the seam waves are substantially reduced in amplitude by the intervening fault. A series of shot-points between 2 and 3 were used to generate reflected seam waves off the fault, detected by geophone spread B. The seismic record in Figure 8(d) shows reflected arrivals on a trend from 0.2 sec. on the left to 0.1 sec. on the right. Ruter and Schepers (1979) note that the reflection is only visible with frequencies above 250 Hz; since the fault is only of seam-thickness in displacement, energy at lower frequencies is partly transmitted through the fault as observed in Figure 8(c). The example demonstrates how conventional filtering and display techniques used in surface seismic exploration can assist in locating faults with seam wave reflection data. IN-SEAM SEISMIC TECHNOLOGY IN THE UX The British approach to development of I.S.S. techniques is unique in that a group has been set up within the National Coal Board with the object of developing instrumentation, data processing and interpretation techniques specific to the particular geometries and wave-types encountered

in in-seam seismic surveys (Buchanan, 1979). Two advances in data processing appear particularly significant. These are the development of wavelet compression techniques and of automated procedures for the production of "fault probability" maps from seam wave data. Wavelet compression is illustrated in Figure 9. The dispersed wavelets are reduced in length by a factor of four, which improves separation of modes or reflected arrivals and hence improves resolution of interpretation procedures. Fault probability maps are produced by lag-sum array data processing techniques. The elementary principles may be understood with reference to Figures 6 and 7. The line R on Figure 7 was calculated from a known seam wave velocity and an assumed fault position. Summation of detected energy in the records along this line could be plotted on Figure 6 at the assumed fault position as a number indicating "apparent reflected energy" or the probability of fault existence. Dispersed train

signal

Figure 9

echo

A dispersed wavelet sequence before and after application of wavelet compression techniques (reproduced from Mason, Buchanan and Booer, 1980 b).

A different fault position would yield a different line R and a different sum to be plotted on the mine map. With all seismic records and survey geometry stored in computer files, this procedure can be automl!ted to evaluate a grid of all possible fault locations and produce a contour map of fault probability. Details of the procedure and different display techniques are given by Mason, Buchanan and Booer (1980 a; 1980 b). Two case histories demonstrate technical and costeffective success with the British I.S.S. methods. Panel Wl 7 in Bold Colliery (Figure 10 (a)) was planned to have a retreating longwall operation from BD. The difference in levels between B and D suggested a structural disturbance within the panel, so an I.S.S. survey was designed to give transmission data from heading CD to AB, and reflection data from CD. Lag-sum processing of

27 the data produced the fault probability contour map shown in Figure lO(b ). The westward extent is clear, but the eastward extent of the fault is not revealed, due to restrictions on survey geometry. Two headings GH and BJ shown in Figure IO(c) subsequently confirmed the I.S.S. prediction of a fault and the mine adopted a shortwall operation from GH instead of the original longwall plan. An example from South Kirkby Colliery couples both surface and in-seam seismic techniques. A surface reflection survey indicated two faults subparallel to and approximately 200 m from heading AB (on Figure l l(a)). An I.S.S. reflection _ survey was commissioned to map the faults pnor to planning a longwall operation south of AB. The result of the survey and its interpretation are shown in Figures l l{b) and (c). Confirmation of the trend of the predicted fault was subsequently obtained with heading XY. A possible geological interpretation is the existence of a set of en-echelon faults the first being too small to be detected from the s~rface. This example illustrates the different but complimentary roles of surface reflection and I.S.S. techniques in colliery location and panel planning respectively.

~·

\ l.

(a )

Wl7

...

11• • t

••

11f113

W15

( C)

200•

CONCLUSIONS The in-seam seismic reflection technique has proved successful in locating faults in the Carboniferous coal seams of Western Europe where seam thicknesses range from 1.5 m to 3 m. Brentrup (1979) quotes costs of A$20 000 to $40 000 for an I.S.S. survey, compared with losses of $0.5 million to $2 million where longwall faces have been abandoned after encountering unexpected faults. In Australia, transmission studies have been successful in a 6 m seam, and with the increasing adoption of longwall mining methods in th~ country_ there is an incentive to develop reflection techrnques suited to local conditions. The use of predevelopment boreholes ahead of headings suggests that hole-to-face techniques will alsp be developed as a planning tool in ·this country. Since presentation of this paper,_two addi~ional field trials have been completed m Australia using face to face geometry in a 3 m seam at Metropolitan Colliery, and face-to-borehole geo~etry in a 3 m seam at App in Colliery. Interpretation techniques suited to data from all three test sites are under development.

1·11• lft- H Oftl Mf1ffik prtd ktioft

w---- - - - - ·-

- --

,.. " · ---z

Wl7 G D

....

Figure 10

(a)

WIS

Bold C.Olliery prior to LS.S. survey (b) fault probability map following LS.S. survey ( c) Bold C.Olliery with headings intersecting faults predicted by LS.S. survey (reproduced by permission of N.C.B., U.K.)

28 (al

_JL._JL

2?""'

IOO

]□[

G20

REFERENCES Barth, G., 1979. Long range in-seam seismics : a case from practice. Gluckauf, 115, p 383 - 384 (English translation). Brentrup, F.K., 1979. Seam wave seismics for proving reserves. Gluckauf. 115, p 381 382. (English translation).

--- --- - - - - - - - - - - -·- ---

,urfoce1el11Y1ic pred lctio11

---- ---- ---- ----

Buchanan, D.J., 1979. The location of faults by underground seismology. Colliery Guardian annual review, year 1979, p 419 - 427. Dresen, L. & Freystatter, S., 1976. Rayleigh channel waves for the in-seam seismic detection of discontinuities, J. Geophys. 42, p 111 129. Freystatter, S. & Dresen, L., 1977. Ausbreitung von Rayleigh-Kanalwellen in Steinkohleflozen - Modellseismiche Untersuchungen. J. Geophys. 43, p 807 - 828. Freystatter, S. & Dresen, L., 1978. The influence of obliquely dipping discontinuities on the use of Rayleigh channel waves for the inseam seismic reflection method. Geophys. Prosp. 26, p 1 - 15.

( C)

_JL._JL

zoom

]□[

GZO A

---- - ,,,.,,...,,,.,,,. --- ------------------------ ---- -

·-

ln -HOffl Hlun lc prfflcttofl

Figure 11

-- -

1urfoc• t1tlm lc pndlcflon

(a) South Kirkby Colliery with faults predicted by surface seismic survey (b) fault probability map after LS.S. survey ( c) Colliery map with heading intersecting a fault trend predicted by LS.S. (reproduced by permission of N.C.B., U.K.)

ACKNOWLEDGEMENTS The authors thank management and miners at Leichhardt Colliery for their assistance in conduct of the trial survey, and thank Dr. D.J. Buchanan of the National Coal Board, U.K., for fruitful discussions and permission to quote two N.C.B. case histories.

Grezl, K.J., Leung, L. & Ahmed, M., 1980. Microseismic monitoring for gas outbursts at Leichhardt Colliery. Symposium on the occu"ence, prediction and control of outbursts in coal mines, p .151 - 158 (Aust. Inst. of Min. and Met., Melbourne). Guu, Jeng-Yih, 1975. Studies of seismic guided waves : the continuity of coal seams. Ph.D. thesis, Colorado School of Mines, Denver (unpublished). Hasbrouck, W.P. & Hadsell, F.A., 1976. Geophysical exploration techniques applied to western United States coal deposits, in Coal Exploration, proceedings of the first international coal exploration symposium (W.L.G. Muir, ed.) Miller Freeman Publ. Inc., San Francisco. Krey, T.C., 1963. Channel waves as a tool of applied geophysics in coal mining. Geophysics, 28, p 701 - 704. Krey, T.C., 1976. In-seam seismic exploration techniques. in Coal Exploration, proceedings of the first international coal exploration symposium (W.L.G. Muir, ed.). Miller Freeman Publ. Inc., San Francisco.

29 Mason. I.M., Buchanan, D.J. & Booer, A.K., 1980 a. Channel wave mapping of coal seams in the United Kingdom. Geophysics, 45, p

1131 -1143. Mason. I.M., Buchanan, D.J. & Booer, A.K., 1980 b. Fault location by underground seismic survey.JEE Proc., 127 Pt F, p.322 - 336. Millahn, K.O., & Arnetzl, H.H., 1979. Analysis of digital in-seam reflection and transmission surveys using two components. Paper presented at the 41st meeting of the E.A.E.G., Hamburg, May 1979. Peterson, S.D., 1979. Modal analysis of seismic guided waves in coal seams, Ph.D. thesis, Colorado School of Mines, Denver. (unpublished). Reeves, J.J., 1979. Investigations of seismic seam waves in U.S. coals. M.Sc. thesis, Colorado School of Mines, Denver, (unpublished). Regueiro, J., 1980. Hole to hole seismic seam wave study in the Zulia coal basin, N.W. Venezuela, M.Sc. thesis, Colorado School of Mines, Denver, (unpublished). Ruter, H. & Schepers, R., 1979. In-seam seismic methods for the detection of discontinuities applied to West German coal deposits, in Coal Exploration vol. 2, Proceedings of the 2nd international coal exploration symposium, Denver, October, 1978. (G.O.Argall, ed.). Miller Freeman Publ. Inc., San Francisco. Su, Fu-Chen, 1976. Seismic effects of faulting in coal seams; numerical modelling, Ph.D. thesis, Colorado School of Mines, Denver, (unpublished).

Bolt - hole and borehole - core logging instrumentation K.G. McCracken and P.J. Mathew C.S.I.R. 0. Division of Mineral Physics, P.O. Box 124, PORT MELBOURNE, VIC. 3207

ABSTRACT A bolt-hole probe and a coal borehole-core analyser have been developed by the C.S.I.R.O. Division of Mineral Physics as part of a major program in the application of nuclear techniques to exploration, extraction and quality control in the mining industry. In some Australian underground coal mines, the exposed roof is incompetent mudstone or shale overlain with sandstone having excellent bolt anchoring characteristics. It is necessary therefore to determine whether the bolt holes have penetrated the sandstone stratum. The bolt-hole probe, known as "Stratasnoop", measures the depth of the mudstone and delineates the boundary between muctstones and sandstone. The device consists of a Csl(Na) gamma-ray detector and associated electronic equipment. The mudstone is detected by its high natural gamma-ray activity caused by the presence of trace amounts of K, U and Th. In the same geological horizon, the sandstone is characterised by its relatively low gammaray activity. The "Stratasnoop" probe is 25 mm in diameter and is provided with a series of fibreglass extension rods. Extensive tests in underground coal mines have demonstrated that the instrument can locate the mudstone/sandstone interface to ±50 mm. A Sydney-based company has been licenced to manufacture the instrument. The borehole-core analyser is a portable instrument for the non-destructive and contactless measurement of density and ash content of coal borehole cores in 5 mm sections. The analyser is based on a gamma-ray attenuation technique and consists of a high-energy and a low-energy gamma-ray source and a Nal(Tl) gamma-ray detector and associated electronics. The borehole core is placed in the path of a well collimated gamma-ray beam from the sources, and the detector measures the attenuation of the gamma-rays. The attenuation experienced by the high-energy gamma-rays is a measure of the density of the bore-hole core section. The ash content is derived from the attenuation experienced by both the high-energy and the low-energy gamma-rays. The laboratory measurements on simulated borehole cores made of coal pellets of accurately known ash content and density showed an accuracy of 1% ash and 0.01 g/cm 3 respectively for coals from the same area. The radiation sources are well shielded in the lead housing. The radiation level in any accessible part of the instrument is comparable to the natural radioactivity from the environment.

INTRODUCTION Research and development in the application of nuclear geophysical techniques to exploration, extraction and quality control in the mining industry is one of the major programs carried out by the C.S.I.R.O. Division of Mineral Physics. The principal nuclear geophysical techniques studied under this program are the natural gamma, gamma-gamma, neutron-gamma anc! neutronneutron methods. Several instruments developed under this program are being licenced to Australian companies for manufacture and sale. A bolt-hole probe and a coal borehole-core analyser have been developed recently by the Division; these instruments form the subject matter of this paper. The bolt-hole probe, which is

based on the detection of natural gamma radiation, is used to determine the location of the interface between mudstone and sandstone in bolt holes drilled for roof support in underground mines. A gamma-ray transmission technique is employed in the borehole-core analyser to measure the density and ash content of coal borehole cores over 5 mm length intervals.

THE BOLT-HOLE PROBE In most underground operations, rock bolts are needed to support the roof of the mine. These steel bolts, which are usually about 25 mm in diameter and several metres long, are inserted into holes drilled in the rock and anchored in rock strata with suitable stress-bearing capabili•

32 ties. In some Australian underground collieries the exposed roof is mudstone or shale of relatively low strength overlain by a sandstone that has excellent bolt-anchoring characteristics. It is therefore important to ensure that the bolt holes have penetrated the sandstone. The thickness of mudstone varies in different areas and there has previously been no accurate way of measuring this parameter. Estimating the thickness of the mudstone by examining chips from the bolt hole is messy, time-consuming and inaccurate. The bolt-hole probe, known as "Stratasnoop", overcomes this problem and can be used to locate the interface between mudstone and sandstone with a high degree of accuracy.

Principle of the Bolt-hole Probe The operation of the bolt-hole probe is based on the measurement of natural gamma radiation (Lynch, 1962). Virtually all geological strata contain naturally occurring radioactive elements such as uranium, thorium and potassium in trace quantities. The presence of these elements can be detected with a radiation detector sensitive to the gamma-rays they emit. The amount of natural radioactivity in a rock formation depends on its composition and geological history. Clay minerals have high absorption and ion-exchange capabilities that result in preferential accumulation of the above radioactive elements. Thus, mudstone and shale, which contain a high proportion of clay, tend to have significantly more natural radioactivity than sandstone, while coal has very little intrinisic radioactivity. The amount of radioactivity emitted from such formations can be determined with the aid of a sensitive gamma-ray detector and associated electronic equipment, and the position of the interface between mudstone and sandstone can be determined from changes in the radiation level recorded. Technical Details of the Bolt-hole Probe The prototype bolt-hole probe consists of a Csl(Na) scintillation detector and photomultiplier 50 mm long x 18 mm in diameter, a high-voltage power supply, an amplifier and a ratemeter. A schematic diagram of the probe and associated equipment is shown in Figure 1. The Csl(Na) detector was chosen for its high efficiency and its resistance to mechanical shock. The photomultiplier requires stabilised power for consistent results, and a high-voltage power supply was specially designed for this purpose. The power supply provides an output of 1300 V, which regulated to ±0.01% over the full range of the battery voltage used (10 V to 14 V). The probe is encased in a 25 mm diameter stainless steel tube. The high-voltage power supply is also incorporated in this tube in order to make the instrument safe to use in underground mines. The probe is powered by lead-acid gel cells that

'•.

Csl(Na) detector

·.·•

.. ...

Photomu lt i pl ier tube Ampl i fier Hi gh- voltage supply

Rate meter

Figure I

Schematic diagram of the bolt-hole probe.

provide 12 hours of operation, and a small battery charger is provided with the unit. The ratemeter, the power cells and the battery charger are enclosed in a leather case that hangs from the operator's neck. The probe has been designed in such a way that a gamma-ray source can easily be incorporated in order to turn the instrument into a gammagamma density probe. Fibreglass extension rods 1.8 m long are used to push the probe into the bolt hole. A series of readings, each over a period of 10 to 15 sec, is taken with the probe at various positions in the hole. A scale on the fibreglass rods indicates the position of the probe at the time of each reading, and the point at which the probe's response changes significantly is taken as the interface between the two contrasting rock types.

Field Test of the Prototype A prototype of the probe was field-tested for six months in order to evaluate its applicability and reliability for routine logging of roof bolt holes in underground mines. Typical logs of selected bolt holes in an underground coal mine working the Bulli Coal ·in the Southern Coalfield, N.S.W. , are shown in Figure 2. These holes had been drilled to a depth of 1.8 m. In holes 106 and 107, the boundaries between mudstone anu sandstone are clearly delineated at a depth of a little over 1 m. The natural gamma radiation level of the mudstone in these holes is nearly twice that of the sandstone. The mudstone extends over nearly the full depth of holes 3~ 5 and 7, but no mudstone is visible in hole 101. Low radiation levels were recorded near the start of holes 99 and 100 because of the presence of

33 10

Ho le 10 6

Mud stone

.... tf .... o-·"'°o

----o- --'Q Mudstone ~

,o a><>--o· Sandstone

0 ' - - -~ - ~-

0 10

0 '---'------''--0 10 Hole 101

______q __ .,,..., __ o

0'-~'-~--

"' C

Sandstone

0 '---'------''--10

il u 0 '---'----''---0 10

0 '---'------'- 0

0 '--'------''--0 2 Depth (m)

Figure 2

hole cores for geological, technical and commercial information. The development of x-radiography of borehole cores (Jones, 1970) has been a significant contribution in meeting an important need in this field. X-radiography provides a permanent record of the nature and distribution of mineral matter in coal in a qualitative fashion. The borehole-core analyser has been developed to determine quantitatively the density and a$ content of borehole cores and broken pieces of coal, either in the field or in the laboratory.

Typical logs of bolt holes.

carbonaceous shale, which has very little radioactivity. However, the remainder of each of these holes penetrated only mudstone. The field tests showed that five or six readings were usually sufficient to locate the mudstone/ sandstone interface to within 50 mm. It should be noted that the probe can identify not only sandstone and mudstone, but also any other interface between rock types that has a characteristic contrast in natural gamma radiation level. The field tests proved the usefulness and reliability of the device as well as the robustness of its mechanical and electronic components. The tests also suggested a few design modifications for improved handling and performance. Such modifications include: (a) incorporation of a digital integrating meter to determine more accurately the probe's responses; (b) redesign of the extension rods and incorporation of a continuous cable on a recoil winder to eliminate the need for switching off the probe when extension rods are added or removed ; (c) reduction of the length of the probe, and provision of a earring pouch. The instrument has now been licenced to a Sydneybased company (ETP Pty Ltd, St Leonards, N.S.W.) for manufacture and sale.

THE BOREHOLE-CORE ANALYSER Very few instrumental techniques have been available to evaluate non-destructively coal bore-

Principle of the Borehole-core Analyser The density and ash determinations are based on measurements of the Compton linear-attenuation coefficient and the photoelectric mass-absorption coefficient of the material (Fushimi, 1968 ; Mathew, 1976). These measurements utilise the attenuation experienced by a "hi_gh-energy"_gamma~ray beam (>300 keV) and a "low-energy" gamma-ray beam (< 100 keV) when transmitted through the coal sample. In a narrow-beam geometry, the attenuation experienced by the high-energy gamma-ray beam is related to the density of the material through which it passes. The density p, is given by a b ln 11 _ p =-d d

Where I l. i_s the intensity of the transmitted beam, d is the thickness of the material, and a, b and c are constants. The ash content of the sample is derived from the attenuation experienced by both the high-energy and the low-energy gammaray beams (Fushimi, 1968) and is given by y =ATr - B

where y is the percentage of ash in the sample, A and B are constants, and T = ln (12flo2) r ln {11/Io 1) where 10 1 and I 1 are the incident and transmitted intensities of the high-energy beam and 10 and 1 are the incident and transmitted intensilies of 2 tne low-energy beam. T is termed the two-beam ratio. r

Technical Details of the Borehole-core Analyser A schematic diagram of the analyser and the associated electronic equipment is shown in Figure 3. The analyser consists of: (a) a 10 mCi 137 Cs source to provide the "highenergy" ~amma-ray beam (662 keV), and a 50 mCi 41 Am source to provide the "low-energy" gamma-ray beam (60 keV); (b) a detector assembly consisting of a Nal(Tl)

Preamplifier

Single- channel analyser

Detector shield and col I imator

Counter Bore-hole core

Single - channel analyser

Counter

Timer

Source shield and collimator

Gamma -ray sources

Figure 3

Schematic diagram of the coal borehole-core analyser.

detector 51 mm x 51 mm in size, a photomultiplier and a preamplifier; (c) a nuclear instrumentation module consisting of a stabilised amplifier, two singlechannel analysers and two timer-counter units; (d) a core rack on rollers for positioning the drill cores in the gamma-ray beam; (e) a programmable calculator interfaced with the timer-counter units. The gamma-ray sources are enclosed in heavy lead shielding, which also acts as a collimator (5 mm wide x 15 mm long) to provide a parallel beam of gamma-rays. The Nal(Tl) detector is well shielded and collimated to accept the gammaray beam. The signals from the detector assembly are fed into the timer-counter units through the amplifier and the single-channel analysers. One anaIyser is adjusted to accept the "high-energy" signals, the other the "low-energy" signals. The amplifier is stabilised on the full energy peak of the 137 Cs source. The programmable calculator determines the density from the "high-energy" gamma-ray count and the diameter of the drill core by means of a calibration equation. The ash content is calculated from the "high-energy" and "low-energy" gamma-ray counts by means of another calibration equation. In the work reported here, every measurement implies a core section 5 mm wide becaus7the width of the gamma ray beam can be vaned to provide the required spatial resolution.

Laboratory Tests of the Borehole-core Analyser

The analyser was subjected to laboratory tests with coal samples having accurately known relative densities and ash contents. The pulverized samples were either put in plastic vials or pe~etized with a hydraulic press to form 50 mm diameter x 50 mm long cylinders to simulate sections of core. The densities of the simulated core sections were determined from their masses and physical dimensions. The coal samples were . collected from nine different seams and locations in New South Wales in order to cover a wide range of ash contents and ash compositions. The ash contents of the samples varied from 6% to 55%. During measurement, each sample was placed on the core rack and positioned in the gamma-ray beam. An intergration time of 100 sec was used to obtain the intensities of the "high-energy" and "low-energy" gamma-ray beams transmitted through the sample. -Counting for the same duration, without the sample, provided the inten: sities of the incident gamma-ray beams. From these two counts, the density and the ash content of the sample were derived. Results and Discussion

Standard regression analysis techniques were used to establish the relationship between the attenuation experienced by the gamma-ray beams and the ash content and density c f the coal samples. The relationship between the ash content and

35 2 .0

l.c

"'E<J ......

E :,

Cl)

-~Cl)

. C

0 3 .0

2 .5

3 .5

Two-beam ratio

Figure 4

1.4

2 .0

/ .. ,..

1.6

> ...... Cl)

••

Cl'

1.8

Relationship between the ash contents and the two-beam ratios of coal samples from the same region.

the two-beam ratio was established with the general equation y = a + bx where y is the percentage of ash (on a dry basis), a and b are constrants and x is the two-beam ratio (T r). Regression analysis of data for 23 coal samples from a single colliery near Wollongong yielded the equation y = -123.980 + 57.186x with a standard deviation of 1.0% ash for a single measurement. The ash contents of these samples ranged from 9% to 55%, and the relationship between the percentage of ash and the two-beam ratio is illustrated in Figure 4. Exclusion of data for high-ash samples {>40% ash) yielded the equation y = - 124.206 + 57 .282x with the standard deviation reduced to 0.7% ash. Regression analysis of data for 41 coal samples obtained from nine different seams and locations in N.S.W. yielded the equation y = -116.501 + 53.59x

with a standard deviation of 2. 7% ash for a single measurement. This increase in standard deviation was evidently caused by the wide variation in ash composition. Thus, the instrument must be separately calibrated for different coals from different areas.

,/'

1.2

•

/✓

1.0 .,__ ___,_ ___.__ __.__ _....,__ __, 1.0

1.2

1.4

1.6

1.8

2 .0

Density (rod ionietric) (g/cm3)

Figure 5

Comparison of density (mass/volume) and density (radiometric) for pelletized coal samples.

The regression equation obtained for the relationship between the density, the gamma-ray counts and the sample thickness was p = 195.12x -

14.25y - 0.025

where x = I/diameter and y = ln (gamma-ray counts)/diameter. The standard deviation of p was 0.01 g/cm 3 for a single measurement. A comparison of the densities of the pelletized coal samples, as determined from their masses and volumes, with the radiometrically determined densities is shown in Figure 5. An example of the ash content and density of a 1 m core section is shown in Figure 6. These logs show the close correlation between the ash content and the density of coal, as well as the distribution of plies and bands in fine detail. Several thin dirt bands, which were not visable in the core section, are clearly visible in the logs. Logs with such details could assist in accurately subdividing cores for destructive testing, seam correlation, coal reserve estimation and information on coal washability. The analyser is currently undergoing detailed studies to evaluate its potential in coal exploration and mining.

Radiation Safety The radiation sources are well shielded in the lead housing of the analyser, and radiation leakage is negligible. The radiation level in any accessible part of the instrument is comparable to the natural background radiation level from the environment.

36 Sydney-based company for manufacture and sale. The coal borehole-core analyser, which is based on gamma-ray transmission, is used for non-destructive and contactless measurement of the densities and ash contents of coal bore-hole cores in 5 mm sections. Laboratory measurements on simulated core sections made of pelletized coal or pulverised coal in plastic vials showed an accuracy of 1% ash and 0.01 g/cm 3 for coal from the same area. The accuracy of the ash determinations decreased to 2. 7% ash for coal samples from different regions because of variations in the mineral content. For accurate ash determinations, coals from different areas need separate calibrations. A sample log of the ash content and the density of a 1 m core section shows the distribution of plies and bands in fine detail . The analyser is currently undergoing detailed evaluation of its potential in coal exploration and mining.

40 E

a. 0•

ACKNOWLEDGEMENTS Messrs D. Coward and B. Maher were responsible for the development of the bolt-hole probe. The assistance of Messrs R.G. Wilson and W.A. Williams of Australian Iron and Steel Pty Limited is gratefully acknowledged. Support for the work on the coal borehole-core analyser was provided under the ational Energy Research , Development and Demonstration Program administered by the Australian Department of National Development.

100 0

20 40 60 80 100 0 Ash(%)

1.0

2 .0

Density ( 9 /cm3)

REFERENCES Figure 6

Sample log of a 1 m coal core length.

CONCLUSIONS Two ½'1struments based on nuclear geophysical te~hniques ~ave been developed for use in exploration and mming by the coal-mining industry. The bolt-hole probe is based on the measurement of natural gamma radiation and is capable of determining the mudstone/sandstone interface in bolt holes drilled for roof support in undergro~nd coal mines. The instrument is simple in design and operation, and it is compatible with the requirements of the .S. W. Department of Mines for use in underground coal mines. Extensive field tests of the probe have proved the usefulness and reliability of the device as well as the robustness of the electronic and mechanical components. Usually, five or six readings of the probe inside a bolt hole will determine the location of the mudstone/sandstone interface to ± 50 mm. The field tests revealed a few minor desig~ modifications that could improve the handling and performance of the instrument. The bolt-hole probe has been licensed to a

Fushimi, H., 1968. Ash content in coal. Rep. NSJ-Tr 134, Japan Atomic Energy Research Institute. Jones, R.H ., 1970. Bore core radiography. Journal of Physics E: Scientific Instruments, 3, p 363 -366. Lynch, E.J., 1962. In Formation evaluation p 243 - 249, Harper and Row, ew York. Mathew, P.J., 1976. Gamma-gamma methods of measurements in boreholes. In Geophysical techniques in borehole applications, (McCracken, K.G., ed.) p 70 - 98 Australian Mineral Foundation Inc. Adel~de.

The uses and benefits of down - hole geophysical logging in coal exploration programs R.I. Renwick The Shell Company of Australia Limited, 155 William Street, MELBOURNE, VIC. 3000

ABSTRACT Down-hole geophysical logs measure the insitu properties of the strata encountered in a borehole from the repsonse to various acoustic, electrical, mechanical and nuclear measuring probes. Such logs are an ideal tool for qualitative geological interpretations and can be used in stratigraphic correlation, location and accurate thickness determination of coal seams, lithological identifications and classification of depositional environments. Quantitive interpretations have also been developed to establish relationships between log readings and various coal quality parameters, as well as sand/shale ratios, porosities, fluid contents and elastic properties in the rocks concerned. Processing and handling of log data is performed by computer using a set of programs designed to evaluate coal logs. Options such as depth plotting, contouring, crossplotting, mathematical calculations, statistical regression and various forms of report presentation are available if required. Illustrations are given of the various log presentations using actual borehole logs associated with coal exploration areas in Queensland and New South Wales.

INTRODUCTION Since 1973, all coal exploration drilling projects carried out by Shell in Queensland, New South Wales, South Australia and Tasmania have included down-hole geophysical logging services. These services mainly involve wireline logging, a process whereby a number qf measuring probes are lowered by cable into the borehole and the response of formation rock to the probe's sensing mechanism is recorded at the surface. Such borehole logs used in conjunction with the traditional exploration techniques of coring and/ or description of chip cuttings, provide a large amount of data from a drilling program. With the integration of these techniques, a rapid and efficient method is available to find out whether areas containing known coal deposits are of reasonable commercial quality and consistent enough to justify a mining feasibility study. Wireline logging provides information on a variety of geological parameters and is an ideal tool for use in coal exploration. It can be used for qualitative geological interpretations such as stratigraphic correlations, location of and determination of thickness of coal beds, lithological identification and classification of depositional environments. Quantitative interpretations have also been developed to establish relationships between log readings and a range of properties such as coal quality parameters, sand/shale ratios, porosities, fluid contents and elastic rock properties. With the development

over the last few years of a second generation of coal logging tools, accurate and reliable log measurements of formation density, hole size, hydrogen index (neutron log), natural gammaray activity, resistivity and sonic travel time are now available. Added to this has been the introduction of an on-site digitisation system whereby the logs are digitally recorded on to cassette to allow speedy evaluation using extensive computer handling and manipulation. With this improvement, certain requirements from coal logging have evolved and these are summarised in Table 1. It is intended in the following sections of this paper to outline the use and benefits of coal logging both qualitatively and quantitatively and examples are presented from data obtained in actual exploration areas held by Shell.

WGGING TOOLS AND THEIR MEASUREMENTS The logging contractor presently employed by Shell in Australia, BPB Instruments Pty Ltd., offers a package of wireline logging tools capable of logging holes in the range of 75 mm 300 mm diameter. The main BPB coal logging tools available at present are listed in Table 2, together with a dipmeter tool that has only recently arrived in Australia. For effective operation of these tools, the boreholes should be water filled, have a stable wall and constant diameter.

LOGGING FOR COAL

A COllll'Uff SUITE OF COAL BOllf.MOU LOOS

FUTUIE

IEIUIIEIEITS

PIESEIT

10 IIEITIFI IIE CUL

IES

I ' IIEITIFI IIE SEI i II TEIS£1 1 SEII I EITS

!El

TI ESTIILISI TIE

IEISITI . UI. I EI TI IS !ALIE , ID ISTIIE

IAII . Sll,.11 . IELI II LE II SI II ILITI rlEII CT ID IS

TIICIIESI, II Ill CUL IIIT , LITIILIII. STIEISTI

IEIIECIAIICIL PAII IE TEIS , FULTS . IIP AI D STIil£ DF IEDDIIS PLUE S

LDI f UILI TI TII ULIT IDIS COU UTE I ISED

USE IF I II ICDIP UTEI, IITEIICT II! IIHHICS

IIILITI

TI IETEIIIIE TIE ltEIIIIIEI

TD PIESE IT DAT A 11 USII LE FDIIIT

-----==- ·---- -- ...... ----· r '

ll TEINITll!S , COIINI

TABLE 1

Requirement for Coal Logging Figure 1

STANDARD BPB LOGGING TOOLS

CCS:

£Of! C-,,..t,_ ,...... , -~.,.,,.lty

Logs

The Coal Combination Sonde Density Logs

11u .,.11r1

-c-a.,. •C..t1.,_

z .,..,,. .. 1,.

,1,. ...... t _ ....... ,_

11o,-1 l.ogifll • ~ =2 a/111.n 111 coal, 9 lllaln in IIOf'l-c:oal atr,1t.11,

TABLE2

A Complete Suite of Coal Borehole

Standard BPB Logging Tools

There are four primary logging tools which can be of use in the exploration for coal bearing formations - the coal combination sonde, neutronneutron, sonic and resistivity tools. The minimum requirement for coal logging would be the use of the coal combination sonde which allows the simultaneous recording of four different log types in the one pass (short and long spaced density, gamma-ray and caliper). A complete suite of coal borehole logs is shown in Figure I. Boreholes are logged from the surface to their total depth with a vertical scale of I :200 (known as general scale). Across coal bearing intervals a scale of I: 20 is used (known as detail scale). All logs are field digitised onto magnetic tape cassettes, and later converted onto nine-track computer tapes, arranged in standard Shell format for further analytical processing on an IBM 360/370 computer.

The two density logs, (LSD and BRD logs with a 480 mm and 150 mm source-to-detector spacing respectively) are used to measure the apparent density of the formations encountered in the borehole. The tool uses a radioactive source (caesium 137) and two gamma detectors to measure the returned particles after collision with electrons in the surrounding strata. All counting rates are processed at the surface and can be related to the electron density of the strata and thus to their bulk density. As the LSD density sonde has a large source-todetector spacing, it permits a deep penetration of the strata by the gamma particles and will give a more realistic measurement of density. The BRD density sonde has a small source-todetector spacing and a shallow penetration. It can resolve thin, layered sequences very clearly, but is greatly affected by adverse borehole conditions and cannot measure accurate densities in badly caved or mud-caked boreholes. The unit of measurement is the standard density unit (SDU).

The Coal Combination Sonde Gamma-Ray Log Gamma-ray logs record the natural radiation of the differing formations encountered in the borehole, caused primarily by isotopes of the potassium, thorium and uranium families. The unit of measurement is the API unit which is the system used exclusively by the oil industry.

The Neutron-Neutron Log Neutron logs are another type of radiometric log

39 which bombards the surrounding strata with neutrons. The sonde consists of a source, which provides a continuous spectrum of high energy neutrons, and a detector sensitive only to low energy (thermal) neutrons. Depending on the atomic structure of the strata the neutrons are either absorbed or transmitted. Due to the complex nature of the interactions, however, it is sufficient to say that hydrogen is the most effective element in slowing down the neutrons and therefore this log is said to measure the hydrogen index of the formation. The unit of measurement is the standard neutron porosity unit (SNPU).

TYPICAL LOG RESPONSE IN A VARIED COAL FORMATION • r u ,11011 - 1n:u1 • 0•

CM.NII

n••t1 / ,n

The Sonic Log

Sonic logging involves the transmission of an acoustic energy pulse (sound) into the strata and the measurement of the returned pulse by two detectors at a fixed distance from the transmitter. This system measures the time taken for the sound wave to travel a fixed distance through the strata. The velocity of the sound waves is dependent on the composition of the grains, the nature of the pore fluid , the pressure, and the proportion of fluid in the grains. The unit of measurement is the microsecond per foot(µ sec/ft).

Figure 2

Typical Log Response in a Varied Coal Formation

The Resistivity Log

The general principle of resistivity logging is the measurement of variations in potential or current intensity with depth, resulting from an electrical current transmitted down the hole. The most useful resistivity log is the focussed electric log, which incorporates a guard electrode to concentrate the current into horizontal sheets, thereby reducing the effect of adjacent beds on the recorded signal. The unit of measurement for resistivity is the ohm- metre (ohm-m).

IIIIPtTC l

rnun: 1 ■

QUALITATIVE EVALUATIONS A combination of the density, gamma ray, resistivity and caliper logs allow the positive identification of any coal bearing strata. As shown in Figure 2, it is easy to distinguish between coal and sandstone, shale or most other lithological units. Lately the neutron log has also been used with reasonable success to help identify more difficult lithologies. Table 3 is a summary of the expected response of the main coal logs to different types of sediments within a coal bearing sequence. Coal beds and individual coal plies can be correlated throughout a deposit with the use of the general gamma-ray or density logs. Figure 3 is a stratigraphic correlation through the Drayton coal deposit, New South Wales. These types of section can be plotted with reference to each individual borehole elevation above standard datum to show the geological structure of the deposit. The detailed BRD log, with vertical resolution

TABLE3

Coal Log Lithology Identification

down to 80 mm, allows the most accurate determination of depths, thickness and sampling intervals. These detailed logs of the coal are very useful when examining the core to verify core sequence, to determine the position of core losses and the nature of the core which may be missing. An example of the use of the detailed BRD log is shown in Figure 4. Cross Plotting Techniques

As the logs record the characteristics of the total interval under investigation, a firm knowledge of how each individual logging tool responds to the various lithologies is critical. By cross-plotting various combinations of logs it is possible to extract still more information on lithology and depositional environment.

40 CROSS-SECTIO N STRATIGRAPHIC LOG CORRELATION

Figure 3

Cross-Section Stratigraphic Log Co"elation

DETAILED COAL SEAII LOO CORRELATION PLOT

Locaton of the coal matrix points is a difficult proce!s from this cross plot. However, a triangle drawn from the sand and shale points to envelop all coal data points gives a good approximation. As the coal seam under consideration is perhydrous high volatile bituminous, the density for the coal point should be between 1.20 gm/cc and I.JS gm/cc. Other ;ross plots shown on Figure 5 are the neutron/gmnma-ray log and the sonic/gamma-ray log. 0 :tee again the various points of interest are lac 1ted to show how this particular combination of logs respond to the sedimentary rocks of the ::tudy area (the Surat Basin). The final cross plot shown in Figure 5 is that of the sonic/ resistivity log. Here again the various lithology points are located. To understand shale resistivity from this cross plot, the interpreter should be conversant with the fact that resistivity increases with decreasing porosity or water omtent. Other influences include salinity changes (resistivity decreases with increasing salirity), temperature changes (resistivity decreasi·s with higher temperature) and nonclay mineral content (resistivity increases with an increase in non-clay minerals). From this type of cross plotting a large amount of information ca11 be extracted to enable further understandin~ of the geology of a coal deposit.

QUANTITATIVE INTERPRETATION METHODS l:!l . ,

□ -·-

il!l " " c::ic::ii::::J "•' l::::J ... ,. □-• □ ~

..·-

D • ••·-

a -·-

With the advent of second generation coal logging tools, it became evident that there is a greater possibility of determining accurate log derived coal quality parameters. Previous studies using the "Mono" type logging tools were not successful due to the non-consistency of the log response, i.e. logging tools of the same type had different responses in the same hole.

Tool Calibration Requirements

Figure 4

Detailed Coal Seam Log Co"elation Plot

The gamma-ray density log cross plot is the easiest of these to interpret. As shown in Figure 5, it can be seen that cleanest sands have a gamma ray value approaching 15 API and 100% shale a value of approximately 90 APL The anomaly known as the gamma marker band is quite clearly visible in the cross plot. A possible explanation of this anomaly is that it represents a shale with a higher percentage of kaolinite as the dominant clay mineral and hence with a greater attraction for water molecules forming between the clay mineral layers.

Prior to any investigation into the use of downhole geophysical logs for coal quality predictions it is fondamental that all factors likely to effect the results adversely are known and quantified. TI.e most sigrlificant effect on the log data will he borehole conditions (e.g. rugosity, washouts, borehole fluid, variations in hole diameter and eccentricity). The next most important factor is the log calibration procedures used on the slin-hole logging tools (e.g. water barrel calibratior,, dead time corrections, statistical errors, accurate conversion oflog readings to density, poroiity etc. and acceptable repeata..bility of response b !tween different logging tools of the same type).

Log Calibntion Techniques For pre die :ion of log derived coal quality values it is recommended that at least three log calib-

OF VARIOUS

EXAMPLES

DENIITY GAMMA RAY

CROSS

LOG

CROSS PLOT

PLOTS

NEUTRON GAMMA RAY CIIOIS

I0OOO T " - - - - - - - - - - - - - - - -..

PLOT

•oo . . . - - - - - - - - - - - - - - - - - - ,

tOOO tOOO

SANO

500 TOOO

.. ·-

COAL ( O. M.M. , . )

►

L.. I

~ ►

5000

iii z •ooo ~

"iiiz

3000 TU,,AC(OUI CLAYITOH[ (GAMM A MUKU )

zooo

100

""' ! U

1000

OUAIIT 0

'•

U.NO/ IH.t.L[

GAMMA - RAY

IIOUNOAllllY

( A"I I

OAMMA - RAY ( AP I)

IONIC IAMMA IIAY CROSS PLOT

SONIC RESISTIVITY CROii

PLOT

1000

INCJIUIINt "0ROIITY ~

CIAMMA - RAY

,000 ..... . - - - - - - , - - - - - - , , - - - - - 1

( A" I l

Figure 5

•o

110

SON IC TRAVEL

Examples of Various Log Cross Plots

TIM[ ( Mt10/fl.)

100

ration core holes be drilled, specifically for the purpose of determining mathematically the possible relationships between log response and laboratory core measurements. It is very important that the core samples selected for laboratory analysis are based on interval units that are clearly distinguishable from the BRD log. Samples are selected to conform to single beds of homogeneous lithology wherever possible, based on continuously sampled lengths no smaller than 350 mm throughout the core. For each of the core samples selected the following calibration testing is usually required: proximate analysis, relative density and specific energy. The resulting analytical data, particularly ash content, can be plotted against depth to produce (say) a core ash log which is then correlated with the LSD log at the same scale as shown in Figure 6. Once the basic correlations are made , the information is entered into a statistical analysis package (SAS) and by cross plotting the data the existence of any trends may be seen. For Shell's holdings in the Surat Basin, correlation between LSD density in CPS and core density in gm/cc is shown in Figure 7. Very good correlations have also been found to exist between ash content and density variations as also shown in Figure 7. CORRELATION LSD AND CORE ASH LOGS

AIM CONTINT

ASH CONTIENT ME ASUR ED ON COAU

* DATA PAI/I$ SELECTED FOR CORE- LOfi CALl.1'A TION

Figure 6

Co"elation LSD and Core Ash Logs _

Further it has been possible to relate the ash content to specific energy and this is illustrated in Figure 7. Because of the many factors influencing these relationships no universal correlations can be made but field wide correlations are available for certain Australian coal-field areas.

i\ccuracy of the Predicted Quality Results To assess how accurately the log derived results itgree with those obtained by laboratory analyses, a selected length of core, not previously used in the calibration, was compared with the results predicted from the geophysical logs. The result of this is shown in Figure 8. From examination of the results of the predicted a.ttd the analysed values for this core hole and others previously studied, it is concluded that the two values agree to within ± 10%. If analysed value for ash content is 15% then the log derived ash content may range between 13.5% and 16.5%.

Further Developments in Coal Logging With the use of the sonic and density logs it is nov, possible to determine a log-derived strength ind1:x, whereby a comparison can be made of the relative strengths of rocks in the sequence. The equations derived to determine the strength index have not yet been compared with known laboratory measurements and at this stage the inde.{ should be used only as a relative indication cif the likely competence of the strata. An example of strength index measurement is shown in Fi1~re 9. Promising results have also been obtained in predictir.g, from the response of the short spaced BRD tool (source-to-detector spacing, 150 mm), the yield upon washing the raw coal at various relati,e densities. Because this log has only recently been redesigned it has not been possible as yet :o compare the results with laboratory float/sink data. An example of log-derived coal predic1ion is shown in Figure 10. COMPUTER PROCESSING AND EVALUATION Becaus1: of the involved mathematical equations used to derive coal quality parameters, Shell has developed an integrated computer system to handle coal core analysis and petrophysical evaluations. This system is based on a data bank of coal core me.1surements and borehole cutting descriptions, COGEO, a log interpretation system, COPLA and a data bank, COPET. It has 'been adapted from the Shell digital log processing and interpretation programs developed over recent years for oil and gas logging. With the recent onsite digiti,ation of geophysical logs onto cassette, much qui cker computer evaluations can be made. All log data cassettes are sent to the logging contractor's 11rocessing centre for conversion to 9

43 LOG CALIBRATION CURVES FOR QUANTITATIVE EVALUATIONS CORE CALORIFIC VALUE v CORE ASH

l!O

c:,

.,".... :I 30 LI.I

::, .J ,c

c:,

...

At:hllt# ,-.,t,,, Nlw • .12·$4 - 0 ·$4$ ( . ,It")

X KMhlcq ltH lln, JM# • Ml·06 -0·.J6/ (a,h" J

LI.I :,:

100

ASH% CORE ASH v CORE DENSITY (AIR DRIED BASIS) IOO

G1 ASH =l-779-2 ·24 (Kentucky)

7 X ASH • l-705-2·011 (Jondaryan)

T e ASH= 1·711-2· 039(ACLANO)

-;;--

oL-----L-'----.......- - - - - - - - - r - - - - - - - - - - r 2 ·0

1·5

1·0

DENSITY gm1/cm 3 CORE DENSITY v LSD LOG 2 .6

DAT A FROM , KENTUCKY JO NOAR YAN

2 .4

SONOE 68 SOU RCE 8 M 2

B. P. B.

2 .2

DENSITY• 6·318 -0-565 /,o ( LSD LOG)

2 .0

.. z

a a

... 1. 6

1.4

1. 2 1000

2000

3000

4 000

0000

8 000

LID LDQ

Figure 7

Log Calibration Curves for Quantitative Evaluations

44 COMPARISON CORE ASH CONTENTS

SURAT BASIN

WITH LOG OERIVED ASH CONTENTS

BOREHOLE STRENGTH DETERMINATION BUI..K WOOULU!

ASH CONTENT

.U I

I00 · 00

DENSITY LOG

SONIC LOG

.U I 11 1

0·00 U .I

Ill .I

I .I

111 .1

CPS

MlCROSEC/FT

PSI

II. I

I.I

I .I

ASH CONTENT MEASURED ON CORES

II .I

ASH CONTENT CALCULAT D FROM WIRELINE LOGS JI . I

U .t

U .I

Figure 8

Comparison Core Ash Contents with Log Derived Ash Contents

track 1/2 inch computer tape. After return to Shell Australia, the tapes are data verified, prior to being converted into a format suitable for storing on the Shell computer based log files and hence allowing access for further processing. Turnaround time from field digitisation until availability on the computer is generally 4 - 6 weeks. Any logs produced prior to fieldsite digitisation can be manually digitised in-house by Shell using a D-MAC. Once the logs are fully digitised, they are stored on an incremental master file. They are stored in depth increments along with the corresponding log readings or deflections. Any corrections or depth adjustments can be made, after which the data can be processed onto a squared master file. This interprets the graphical log trace in a block form. Smooth curves are reshaped into blocks with square angles, each block representing a layer. Once all the logs are in the squared form for each bor.ehole they are merged onto the merged master file. The merging procedure consists of establishing a reference log for depth control, thereby giving one fixed depth reading per layer, plus readings for up to a maximum of nine logs.

U .I

..., Ill .I

I .I .U I ul

Figure 9

H .I

Ill . I

I.I

.U I 11 1

Borehole Strength Determination

A coal log tvaluation program (COPLA) will transfer data from the merged master file and begin analy;ing the coal logs. A flow chart of the comple-:e computer evaluation system is shown in Figure 11. All log illustrations produced for thJ s paper were made using the computer system outlined above. COPLA has been designed to allow various empirical calculation sub-routine; to be used to convert geophysical logs to precicted coal quality. Average and cumulative propertie:, , .m also be calculated using

45 MINEA■LE COAL ,_DICTIONS FROII ■ftD DENSITY LOG

• l.45v/cc • 30% ASH INPUT CONDITIONS : Yie-ld tf washed ot 1.,s glee . 20,. >25 ...

Min.ffilf'W!Qbl• ooc,a t"-:knf515

Mn. mirwoble dn tt-«llness

-;!'

-:('

> Sm > 2m

> 5/10m > 2m

the raw log results. It is possible to determine coal intervals and toal mineable coal for various ash% cut-off values. By processing all borehole logs for a particular coal deposit, a complete case study can be made. Examples of some of these evaluation techniques via COPLA are shown in Figures 12, 13, 14 and 15. All log derived data can then be stored on the petrophysical data bank (COPET), together with the geological data bank (COGEO) containing, next to core analysis data, a fairly complete geological review of each borehole. This provides an integrated geological/petrophysical data bank from which a number of further geological or mine planning studies can be made with the appropriate computer programs.

ACKNOWLEDGEMENT The author wishes to acknowledge the valuable contribution of D. Humphries to the company's coal petrophysical computing system. The paper is published with the permission of the Shell Company of Australia Limited and Shell International Petroleum Company.

CORE ASH

ROENSI LOG OER1YEDASH

~ EC.OIOC:ALLY -.,.BI..E COAL ~DIRT

Figure 10 Mineable Coal Predictions from BRD Density Log

DEPTH PLOT CORE OATA

GEOLOGICAL DATA

Figure 11 Flow Chart - Computer Petrophysical Analysis for Coal Log Data

COMPUTER PRINTOUT LOG QUALITY RESULTS

INPUT DATA LSD PLY SIZE COUNT M /SEC

PLY NO.

CALCULATED VALUES CO~IPOS ITES PLIES NOM. TOP f'lOIS S.E. .SH A . S.G. S.E. THICK DEPTH S.G. ASH M

MS/KG

2.02 69.7 2.00 68.6 1.89 61.6 1.97 66.8 1.99 67.9 1.67 45.9 1.79 54.9 1.76 52.9 1.87 60.6 1.90 62.5 1.93 64.4 1.97 66.9 2.01 69.3 2.03 70.4 1.65 44.1 1.88 61.0 1.85 59.6

4.70 5.11 7.67 5.78 5.36 13.48 10.14 10.89 8.05 7.35 6.65 5.73 4,85 4.43 14.13 7.89 8.43

9.0 9.0

o.o DIRT 0.0

DIRT

9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0 9.0

0.0

DIRT

0.0 0.0 0.0 0.0 0.0 0.0

0.0 0.0

DIRT DIRT DIRT DIRT DIRT DIRT DIRT DIRT DIRT DIRT DIRT

21.48

9.0 9.0 9.0 9.0 9.0 9.0

0.0 0.0 0.0 0.0 0.0 0.0

COAL COAL COAL COAL COAL COAL

0.0

DIRT

16.09

9.0 9.0 9.0 9.0

0.0 0.0

DIRT DIRT

o.o COAL o.o COAL 0.0

COAL

27.43

9.0 9,0 9.0 9.0 9.0 9.0 9.0 9.0 9.0

0.0

DIRT

9.0

DIRT

12 13 14 15 16 17

29.99 30.02 30.29 30.56 30.71 30.98 31.37 31.55 31.76 32.03 32.12 32.24 32.39 32.51 32.84 33.23 33.44

0.33 0,18 0.12 0.12 0.21 0.18

4165 4852 4218 4120 3869 4663

18 19 20 21 22 23

33.77 34.10 34.28 34.40 34,52 34.73

1.45 1.39 1.45 1.46 1.48 1.40

25.6 18.1 25.0 26.1 29.0 20.1

20.95 23.71 21.17 20.76 19.70 22.97 1.14

0.24 0.18 0.12 0.24

3544 2662 2853 3015

24 25 26 27

34.91 35.15 35.33 35.45

1.52 1.65 1.62 1.59

32.8 44.1 41.5 39.4

18.29 14.13 15.08 15.86 0.78

0,36 0.57 0,18 0.48

4940 6098 6019 6077

28 29

35.69 36,05 36.62 36.80 37,28

1.38 17.2 24.05 1.29 5.5 28.37 1.29 6.2 28,09 1.29 5.7 28.30 1 ;i:n 7.0 27.82 1.95

37.64 38.33 38.48 38.60

2.13 2.00 2.05 2.21

;i:c; n \J ,JU

en,,,. JJ,,

0,69 881 0.15 1166 0.12 1037 0.36 723

31 7:'l

J<.

33 34

35 36

0.03 1124 0.27 1177 0.27 1534 0.15 1265 0.27 1209 0.39 2533 0.18 1928 0.21 2056 0.27 1591 0.09 1487 0.12 1385 0.15 1258 0.12 1143 0,33 1090 0.39 2661 0.21 1567 0.33 1650

1 2 3 4 5 6 7 8 9 10

CALC. VOLS.

.J.,JU

75.3 68,8 71.6 79.5

MS/KG

c,o ~.o

1.44 24.2

1.59 38.8

1. 3]

8.0

2.63 5.03 3.99 1.10

Figure 12 Computer Printout of Log Quality Resu.lts

o.o DIRT o.o DIRT

o.o DIRT

o.o

0.0

o.o

o.o DIRT

o.o COAL o.o COAL o.o DIRT

o.o DIRT

SHELL COAL EXPLORATION BOREHOLE VERTICAL SELECTION AND CALCULATED OVERBURDEN RATIOS FROM LOG ANALYSIS

FOR INTERVAL

29 . 99 - 38 . 96 M.

VERTICAL SELECTION IS BASED ON A MINIMUM SEPARATELY MINEABLE DIRT PARTING OF 0.50M THICKNESS AND A MINIMUM ECONOMICALLY MINEABLE COAL THICKNESS OF 0.50M. DIRT IS GENERALLY DEFINED AS MATERIAL HAVING A DENSITY GREATER THAN 1 . 49 G/CC BUT IT CAN ALSO INCLUDE SOME THIN COAL LAYERS. OVERBURDEN RATIO IS DEFINED AS THE RATIO OF THE OVERBURDEN PLUS DIRT PARTINGS TO THE TOTAL MINEABLE COAL. YIELD IS BASED ON COAL BEING WASHED AT 1.49 G/CC. MINIMUM ECONOMIC YIELD IS ESTIMATED TO BE 80%. WORKING SEAM NO. 1 2 3 4

TOP

BOTTm,

33. 77 34.91 35 . 69 37 . 64 -

34.91 35.69 37.64 38.96

THICKNESS

COAL OR DIR"

AV. S.G .

COAL DIRT COAL DIRT

1.44 1.59 1.31 2 . 13

1.14 0.78 1. 95 1.32

GLEAN COAL 1.49 %

S .G. CLEAN COAL

RAW ASH CONTENT %

NOMINAL MOISTURE

100.0

1.44

24.3

100 . 0

s. r .

NONCOAL

9.0

21.45

o.o

27 . 39

9.0

8.1

1.31

AV. RAW S".E. MJ/KG

o.o

COAL IN

DIRT %

OVERBURDEN RATIO VOL/VOL VOL/WT CU.M/CU.M CU.M/TONNE 29.6

20 . 6

11.2

8.2

0.0

SUMMARY MINEABLE COAL TOP COAL

BOTTOM COAL

TOTAL THICKNESS

NO.COAL BENCHES

33. 77 -

37.64

3 . 87

THICK- SPECIFIC GRAVITY NESS 3 . 09

1.36

MINEABLE RAW COAL guALITY (A.D.) MOIST- S.E. SPECIFIC ASH MJ/KG CON- URE. GRAVITY TENT

TONNAGE /SQ.MTR

25.06

4.19

TONNAGE /SQ.MTR

COAL %

4.19

100.0

1.36

14.5

9 .0

COAL/WASTE RATIO M/M. CU.M/T 11.2

8.2

WATER FILLED PART OF BOREHOLE

""'

30/30/50% ASH

CUT OFF

FIELD SUMMARY OF THE COMPUTER ANALYSIS FOR THE DETAIL LSD DENSITY LOGS MI NEABL E COAL TOP

9OTTOM

TOTAL

NO.

__c_q_~L__

THI CK _N_E_~~-

COAL BENCHES

l-0.E CX)llL

? ;;-i .......

"""

--------

MINEABLE_RAW_COAL_OUALITY_(MOISTURE _ BASI S)

-------------

COAL / WA STE

THI CK- SPECIF I C TONNAGE NE SS GRAV ITY ~?9~'!1I!3_

--------

COAL

-~--

SPEC I FIC GRAV IT Y

--------

~9'::II~'::l:r

39. 99

16. I 0

I . 70

1 . 57

2.67

100 . 0

1 . 57

41. I

5.0

17 . 44

2 . 67

22 . 5

21 .I 9

61 . 89

40.70

2 . 20

1 . 58

3.47

100 . 0

1 . 58

41 . 6

5.0

17.25

3.47

2 7. I

1 7. 2

66 68

34.59 -

86. 2 9

5 1 . 70

2. I 0

I . 54

52.29

0.90

1 .-64

3.23 1. 48

100.0 100 .0

1 . 54 1 . 64

38. 1 47. 2

5.0 5.0

18 . 54 15. 22

3 . 23 1 . 48

40. I 57. 1

26.0 3 4 .8

1 5. 99 -

17.89

I . 90

1. 54

2 . 92

100.0

I . 54

37.7

5.0

18 . 67

2 . 92

8.4

5.5

37.79 32.39

59. 1 9 46.49

2 I . 40 14 . 1 0

5 5

3.90 2 . 70

1 . 55 1 . 56

6.03 4.22

100.0 100.0

I . 55

5.0 5.0

I 8. 36

1 . 56

38.6 40. 1

17 . 79

6.03 4.22

1 4. 2 1 6. 2

9.2 1 0. 4

87. 19

10.50

2.50

1 . 52

3.80

100.0

1 . 52

36.2

5.0

i 9 . 22

3.80

33.9

22.3

46.09

5 .00

2. 1 0

1. 51

3 . 18

100.0

1.51

35.6

5.0

1 9. 44

3. 1 8

2 1. 9

1 4. 5

7.20 7.40

4 7

1 . 80 3 . 30

1 . 60 1 . 55

2 . 87 5. I I

1 00.0 100.0

1 . 60 1 . 55

43. 1 38.9

5.0 5.0

16 . 69 1 8. 23

2.87 5 . 11

30.6 23 . 6

1 9. 2 1 5. 3

40.0 46.4

5.0 5.0

17 . 83 1 5 . 52 1 9. 67

12 . 02 3.25 6. 18

4.6 10 . 2 6.9

3.0 l 1.3 4.i5

51. 39

~ q

73 75

43 .09 -

~ b

76 78

49.59 73.89

56.79 81 . 29

80 81 82 84 85 86 87 88 89 90 91

17. 29 2 9.49 -

43.29 40.49

26 . 00 11 . 00

12 3

7.70 2 . 00

1 . 56 1 . 63

1 2. 02 3. 2 6

100 . 0 100.0

I . ~6 1 . 63

2 1 . 99 I 4. 15 1 2 . 83 2 2. 88 38 .90 34.34 -

32.59 33 . 50

10.60

4. I 0

1. 51

10 0.0

I . 51

19.64 3 1 . 73 46. 1 9 53.09

6 . 18

•

6. 81 8.85 7 . i9 18 .75

3 5 1i

68 . 36 37.49 56. 2 1 46.52 48 . 86 31. 28

1 4 .64 27 .3 3 19.38 23 .07 6.75 6 . 45 17 . 52 16.86 10.29

5 9 5 5

96 97 98 1 00

29 . 57 39.89 2 6.00 45.29 30 .74 49 . 76 2 9.00 32.00 20. 99 -

4 4. 2 1 67 . 22 45.38

101 102

20. 1 2 23.87

45.08

39. 1 4

;:i

t-,

fC t-,

:;i;, ~

i:::

.... "'

l:!-!.z.~~

TONNAGE RATIO _/SQ . MTR _ M/ M_ CU. M/ T _

23.80

S .E.

i:::

i::' "" ~

MOI 5TURE

76.69 -

ASH

92 95

5.0 J

":I. VO

1 . 47 I . 56

30 . 9 40.9

5 .0

2 1 . 16

2. 4 i

5. 0

1 . 46 1 . 52

30 . 5 36.6

5 .0

100. 00 97.9 100.0 93.2 100.0 100.0 1 00.0 10 0 .0 100.0

1 . 48 1 . 53 1 . 54

3 2. I 37.0

5 .0 5.0 5.0

5. 14 5. 0 1 11 . 39 6.08 23.66

I . 52

33 .- 6

1 . 44 1 . 45 1 . 45 1. 52 1 . 44

27 .9 28 .9 32 . 1 36. 1 32-:--9

5.0 5.0 5.0 5.0 5.0 5.0

1 7 . 50 2 I . 28 I 9. 07 20 . 7 1 18 . 91 I 8. 63 i O. 1 7 22.2 4 2 I . ii? 2 0. 71 I 9. 2 6 H.4 4

.. - .

1 00.0 91 . ? 1 00.0

•

1 . 65 3.57 3. 1i 2 7.74

1 . 47 1 . 57 I . 1ii5 I . 48

3 4 6 6

4. 1 1 1 5. 81 3.84 6. 1 2 3.03 4.47 7. 7 1 4.23 3.99

1 . 56 1 . 44 1 . 45 1 . 45 1 . 52 1 . 44

6.08 2 4. I 8 5.90 9.52 4. 36 6.48 11 . 2 0 6.43 5.74

2 4 . 96

12 . 6 0

1 . 45

18.31

96.7

1. 4 5

2 9.4

5.0

21 . 71

15 . 27

7. 5 g

1. 59

1 2. 06

1 00.Q

I 5B

42.6

5. Q

16 88

l . Si 1 . 53 1. 54

2.42 S . i5 I 5. 0 I 11 . BO

35.0 ....... v

-.. . ...

96 . 5

31 ·:·B

5.0

5.90 B.B? 1i. 35 o.1iB 11 . 2 0 o.1i:l 5 .74

14.3

• .o I O. 9

~ -V

7.9 12 . 5

5.0 8.S 3.B

5.9 1. B 3.3 10 . 6 I O. 2 , , . 1i

?.1i

5.5 ~. I ?.O

i5. B

6.B ?.O B.O 3.5 i5. 5 1i.B

17 . 70

2 .6

1.8

1 2. 05

4, Z

z 6

11 . 5.0

,o.o

49 GRAPHICAL PRESENTATION OF LOG RESULTS SCALE

1, 200

GAMMA RAY LOG GAMMA

CPS 100

-=c_

SCALE

LSD DENSITY LOG

LONG SPACED DENSITY LOG

DENS ITY CPS

6000

I , 20

LOG DERIVED ASH%

100

~ 30

'A' SEAM

'e' SEAM

'c' SEAM INTERVAL

THICKNESS

AVERAGE ASH%

AVERAGE CALORIFIC VALUE MJ/t<Q 23 . 20

2 . 58m

19 . 5

1. 2m

16. 8

24 . 19

I . 56m

20 . 7

22 .75

4 5

Figure 15

Graphic Presentation of Log Results

BIBLIOGRAPHY Bond, L.O., Alger, R.P. & Schmidt, A.W., 1971. Well log applications in coal mining and rock mechanics, AIME Trans, 250 ( 1971), 355. Burk, J.A., Campbell, R.L. & Schmidt, A.W., 1969. The litho-porosity cross plot, SPWLA 10th Annual Logging Symposium Edwards, K.W. & Banks, K.M., 1978. A theoretical approach to the evaluation of insitu coal, Canadian Institute ofMining Bulletin, April, 1978. Elliot, R.E. & Strauss, P.G., 1977. A classification of coal measures rocks based on quartz content, Compte Rendu de

Congres Intern Strat. Geol. Carboni[. Sheffield, U.K. Vol. II - 1970 - p 715 - 724 Geetz, J.R., Prins, W.J. & l.ogar, J.F. , 1977. Reservoir delineation by wireline techniques, The Log Analyst, Sept - October, 1977. Geyer, R.L. & Myung, S.I., 1970. The 3D velocity log: A tool for in-situ determination of the elastic moduli of rocks,Proc 12th Symposium on Rock Mechanics. University of Missouri November, 1970. Hedlop, A., 1974. Gamma-ray response of shaly sandstones, SPWLA 15th Annual Logging Symposium (G).

50 Jenkins, J.C., 1969. Practical applications of well logging to mine design, SPE A/ME (1963) Kempter, E.H.K., 1968. Guide for lithological interpretation of well logs, Shell U.K. (SEPE/1) September, 1968. Kowalski, J.J., 1975. Formation strength parameters from well logs, SPWLA 16th Annual Logging Symposium (N). Kowalski, J.J. & Holter, M.E., 1975. Coal analysis from well logs, SPE 5503 ( 1975) AIME, Dallas, Texas. l.avers, B.A. & Smits, L.J.M., 1976. Recent developments in coal petrophysics, SPWLA 17th Annual Logging Symposium, Denver, Colorado. Peeters, M. & Kempton, N.H., 1977. Wireline logging for coal exploration in Australia, The Log Analyst, May - June, 197Z Poupon, A. & Gaymard, R. 1970. The evaluation of clay content from logs, SPWLA 11 th Annual Logging Symposium (G ). Poupon, A., Clavier, S., Dumanoir, I., Gaymard, R., & Misk,A., 1970. Log analysis of sand shale sequences - a systematic approach, Journal of Petroleum Technology, July, 1970. Reeves, D.R., 1978. Some improvements and developments in coal wireline logging techniques, Proc. 2nd International Coal Exploration Symposium, Denver, Colorado, U.S.A. Samworth, R., BPB Internal Note. Schlumberger, 1972. Log interpretation principles. Tixier, M.P., Loveless, G.W. & Anderson, R.A., Estimation of formation strength from the mechanical properties log. Weaver, C.E., 1960. Possible uses of clay minerals in search for oil, Bull A.A.P. G. Vol. 44, No. 9 (Sept. 1960). Weltz, L.S., 1976. Electrical logging applied to the Carboniferous seams of the Pecket Zone, SPWLA 17th Annual Logging Symposium ( K). Wtlterbrood, R.T., 1978. Quantitative evaluation of coal composition using density, sonic and natural gamma-ray wireline logs, EP 49065.

The application of geophysical borehole logging to coal exploration Brian Groves and Edward Bowen AMAX Iron Ore Corporation, 55 Maquarie Street, SYDNEY, N.S. W. 2000

ABSTRACT Geophysical borehole logging has now become an integral component of many coal explor. ation and evaluation programmes in Australia and assists in providing a rapid and cost effective means of assessing a deposit. Tangible advantages are afforded by both conventional analogue and digital recording. Analogue recording enables lithological identification, assists the control of drilling operations, and serves as an aid in correlation between boreholes. Digital acquisition and recording has recently become widespread within the industry and provides high quality data suitable for computer processing and quantitative interpretation. The current trend towards acquisition systems based on a mini-computer also enables onsite processing and plotting of geophysical and geological data. Estimates of coal quality parameters such as ash content, relative density and specific energy, can be readily calculated from digitally recorded data to supplement, but not replace, laboratory coal quality determinations. Discrepancies often exist between the laboratory and geophysically determined parameters due to inherent differences in the methods of determinations, and the nature and size of the sample. The geophysical estimates, although generally regarded as the less reliable, are of considerable value in detennining quality parameters in programmes relying heavily on non-cored boreholes. Examples from exploration programmes in New South Wales and Queensland illustrate the value of such digitally recorded geophysical logs in increasing the knowledge of the lithology and coal quality in each deposit. Further efforts are warranted however, to improve the resolution and count statistics of slimline logging sondes and to speed the d~velopment and utilisation of tools such as the sonic, focussed resistivity and neutron activation technique.

INTRODUCTION Since January 1976 the coal division of AMAX Iron Ore Corporation has geophysically logged all its coal exploration boreholes in Australia. A large percentage of these bores has been drilled by open hole methods. In conjunction with logging data, such techniques provide a means of rapid and cost-effective geological assessment. All downhole data was initally recorded in analogue fonn, but at the instigation of AMAX, a downhole logging system with digital acquisition and recording capabilities was constructed in early 1977 by Geoex Pty Ltd of Adelaide. Basic system specifications, defined by AMAX were for a suite of logging tools comprising caliper, single point resistivity, neutron, gamma ray and dual spaced density as well as the necessary winch and surface electronic equipment. Digital recording was to be to industry standard using a recognised code (ASCII) and on open reel 7 or 9 track magnetic tape. Further developments of this system were the incorporation of a mini-computer and plotter to allow on-site data processing. Since

the advent of the Geoex system, several other contractors have introduced similar types of facilities. The geophysical data obtained has been extensively employed for both qualitative and quantitative determination of lithological information and coal properties. There are, however, limitations imposed by the capability of present logging tool and processing methods.

LOGGING TOOLS AVAILABLE A suite of logging tools comprising caliper, dual spaced density, single point resistivity, gamma ray and neutron-neutron has been found to be adequate for most coalfield applications. These tools are readily available in Australia and have applications in lithological identification, drilling operations control and detennination of coal quality. Other tools, including sonic, focussed resistivity and temperature equipment are also becoming more widely available and have app· lication to both qualitative and quantitative data manipulation methods. Most borehole logging contractors utilise four

52 wheel drive vehicles to transport the tools and associated electronic equipment together with the winch, cable and pulley system used to lower and raise the tools in the borehole.

DATA ACQUISITION Geophysical borehole logging data has traditionally been presented in analogue ( conventional chart) form, but recently, with advances in digital data acquisition technology, it has become more common to record in digital form as well. Digital Recording and Storage of Geophysical Data The signals ge11erated by the logging tools may be in one of two possible forms, The data from the electrical logs is in an analogue form, while the radiation, temperature and caliper instruments generate electrical pulses proportional in frequency to the response of the tool. Both these representations of the data can be readily prepared for digital recording. The pulses are counted over an accurately known time period, while the analogue data is converted by an analogue-to-digital device. The measured pulse rate is sampled at known intervals (common sample intervals are 100, 200 and 250 milliseconds) and digitally recorded. Ideally, the geophysical data should be stored on a readily available and easily usable medium. Open reel nine ( or seven) track magnetic tape is especially suitable since it lends itself both to external processing and processing on-site. Recording of data on such a format has to conform to industry standards and be in a recognised computer code. Digital data acquisition and recording equipment is commonly incorporated into the system within the logging truck. An essential requirement of the digital system is for it to have operational reliability under harsh environmental conditions. For this reason, facilities must be available within the logging truck to check recorded data immediately after a borehole has been logged. Ideally, computing facilities should be available within the logging truck for the processing and quality control of digitally recorded data. Several contracting companies offer processing facilities within their logging trucks. With associated on-line plotting equipment, the computing system enables the plotting of expanded scale logs and data processing at the borehole. Advantages of Digital Recording Digital recording enables more information to be obtained from the logging tools. The advantages of digital acquisition are : (a) more efficient handling of data from a large number of boreholes; (b) provision of a detailed expanded scale (1 :50) density log over coal intersections without the need for an extra logging run in the hole at a slower speed. The normal depth scale for log presentation is I :200

(c)

( d)

(e)

but 1)gs can be presented from digital data at vittually any scale desired; provision of high quality data for computer analysis. Even though quantitative estimates ,,f coal quality can be derived from analc ,gue data, the calculation of such estimates is time-consuming, tedious and less acurate. Digital recording and processing e :1ables the highly efficient, automated calculation of quantitative information; on-si te processing of digital data has enabled a same day turn-around on quality estimates from down-hole logs. This can prove invaluablt in planning exploration drilling; digital recording improves the effective resobtion of the logging tools.

QUALITATIVE APPLICATIONS OF GEOPHYSlCAL BOREHOLE LOGGING DATA Perhaps the most valuable qualitative application of geophysical borehole logging is that of lithological interpretation. For interpretation to be successful, the suite of logs run should include caliper, density, single point resistivity, natural gamma ray and neutron. With this combination of tools, it is possible to identify most lithologies with reasonable certainty. lithologicd Interpretation Geophysical log interpretation is an important adjunct to the open hole drill cuttings record. In cases of poor cuttings return or contamination and delay in cuttings return to the surface from deeper holes, log interpretation yeilds more accurate information on lithology and depth. Log responses for each lithology commonly encountered in sedimentary sequences are well defined. Figure I depicts a typical suite oflog responses for a given borehole, together with the interpr,:ted lithologies. By understanding the basic o;Jerating principles of the logging tools, lithological explanations of variations in the typical geophysical responses can be made. A common example of such a variation is the changing resistivity and natural gamma ray count rate assocfated with the variable clay content of sandstones. Correlation Geophysicitl logs provide a continuous and objective method for field-wide coal correlations. The objectivity of the method is emphasised by the subjective nature of bore core descriptions which may be compounded, in some instances, if more than one gi ,ologist performs the core logging. Correlations between boreholes are made on the basis of comparison of the geophysical log characteristics of coal beds and other marker beds. Density lo~ characteristics are generally the most useful for correlating coal beds, while for other strata, responses from the neutron and resistivity logs may prove more important.

53 been investigated in an attempt to obtain cheap, accurate coal quality information from geophysical logs. The prime emphasis has been on the estimation of ash content, but other parameters, such as volatile matter, fixed carbon, moisture content, float fraction yield, relative density and specific energy have also been estimated, each with with varying degrees of success. With the exception of thickness, none of the above properties can be precisely related to the responses of any logging tool, nor are there unique tool responses for any particular quality parameter from area to area. This is due to the complex chemical nature of coal and restrictions imposed by the resolution of the logging probes. However, quantitative data obtained from geophysical logging may still be beneficially integrated into any exploration program, provided the limitations of its accuracy are well understood. The following sections briefly describe the method of estimating each of the above quality parameters, the accuracy of this estimation and some possible reasons for variations in accuracy. Prospect areas near Boggabri in nqrthern New South Wales and Toowoomba in south-eastern Queensland were used for research and experimentation on these estimation methods.

Ash Content Figure 1

Common litho/ogies and their geophysical log responses

Drilling Operations Control Analogue records can also be used to check driller's depths, core losses and sampling intervals. If core recovery in coal is less than the agreed minimum percentage, the geophysical logs can be used to assist in determining where the core loss occured, and thereby possibly prevent a similar situation in any subsequent redrill.

Roof and Floor Conditions Reeves (1971) reported the use of geophysical logs to determine roof and floor conditions of possible mineable units. The principle used was that sections of the borehole wall which collapse or cave are weak and those that stand are strong. The degree of caving and the thickness of the affected section can be readily determined from the caliper log, while classification of roof or floor conditions is obtained by interpreting the short spaced density log.

QUANTITATIVE APPLICATIONS OF GEOPHYSICAL BOREHOLE LOGGING DATA Since the initiation of digital downhole logging in AMAX prospect areas, many possibilities have

Previous literature has described two major techniques of estimating ash content from geophysical logs: conversion from the gamma log and conversion from the density logs. These conversions involve the process of log calibration, that is, determination of the relationship between geophysical log readings and core analysis data. Reeves (1971) described a method of calibrating ash content with natural gamma counts. The radiation level of the roof and floor of a coal bed was assumed to represent a "rock" of 100% ash content. Zero ash was taken as a level equivalent to 10% of that value, with the proportion of ash content following a linear relationship between these points. However, there are several problems that arise from such a simple model. These are: (a) trace elements of relatively high radiation output can exist within coal, thus causing anomalous estimates of ash content; (b) low count rates obtained in gamma logging (less than 50 counts per second) give rise to high statistical variations, especially in the vicinity of bed boundaries; (c) gamma radiations from roof or floor materials distort the readings from relatively ash-free coal at the top and bottom of the coal bed. For these reasons, an ash-natural gamma log calibration in AMAX's experience has not been successful.

Ash estimation from density log data has been attempted worldwide in recent years. Kowalski and Holter (1976), Lavers and Smits (1976), Peeters and Kempton (1976) and Reeves (1971) have reported the results of their work in this area. The success in generating synthetic ash profiles depends largely on the type of coal investigated. As coal is a composite substance, there are many variables which influence ash determination as a function of density. These are: (a) the degree of coalification or rank of the coal being analysed; (b) the type of mineral matter associated with the coal (whether it is clay, quartz, pyrite, carbonate); (c) the maceral composition of the coal; (d) the moisture content associated with either the mineral matter or the coal substance.

Method of Log Calibration for Ash Determinations Several cored holes should be selected for calibration purposes at the beginning of a field program. Any coal encountered is sampled in units exhibiting homogeneous density log characteristics. Wherever possible, these samples are also selected to coincide with units of uniform lithology. They are analysed in the laboratory for raw coal ash and relative density and the resulting analytical data (ash content) are plotted against the appropriate geophysical log, such as the high resolution density or long spaced density log. The data pairs are also processed through regression analysis programs to find the best possible calibration relationship. In AMAX's experience, it was found that third-order polynomial nomial relationships produced optium correlation coefficients. Data Processing The process of ash determination from rawdensity tool data is simplified markedly if processing facilities are available in the field. By incorporating the ash-density count rate relationship into suitable computer programs, the raw density count rate is read from the magnetic tape and converted to ash content estimates. Presentation of the ash estimates are in the form of an apparent profile over the required coal intersection (Figure 2). RESULTS AND ACCURACY In the techniques used for the Boggabri area, N.S.W., ash content can be estimated to an absolute accuracy of± 4%, a degree of precision that would be acceptable in a high ash coal. However, since the ash content is generally less than 10% (air dried), such estimates only achieve a relative accuracy of approximately ± 40%.

ASH PERCENT l )~--2~ □ --4~□--6~□--8~0_ _ _!.00

DEP H (M l

Figure 2

Computed ash profile, Boggabri area

In the are~ near Toowoomba, Qld., an absolute accuracy c f ± 8% has been achieved for raw ash estimates. The ash content of the coal is higher in this area and hence the relative accuracy of estimation is approximately ± 20%. Because the density pr,)bes measure the bulk density of all material smrounding the sonde, there are many factors which influence ash determinations as a function of density log calibration. These can be divided into physical and chemical factors and are di::cussed below. Physical F 1ctors Affecting Accuracy (a) Witll the technology presently available, it is very difficult to determine adequately the i ize of the borehole cavity in the actual direction of the density measurement. ThrE e-arm caliper tools measure the average borehole diameter only and do not provide sufficient information for valid density log com:ctions. For this reason, a reliable relationship between borehole diameter and vari2tion of density log count rate was not established. It would appear that this problem has not yet been resolved elsewhere in the world. A pc,ssible solution involves the detection in discrete energy bands of the energy

(b)

spectrum of gamma rays emitted by the density tool source. Present density tool detectors measure energies in a continuous range from lOOkeV to 2.6MeV. Figure 3 illustrates the responses of these tools as a function of relative density, with the energy bands corresponding to coal, mineral matter impurities and cavities superimposed. A density tool either employing several discrete energy band detectors or a conventional dector with associated energy bandpass filters should help overcome the cavity problem. The most significant problem encountered in quantitative ash estimation is the effective resolution of the density tools. Theoretically, the minimum resoTuITon of a probe is equivalent to its source-to-detector spacing, in which case a "full peak" response is registered opposite a bed of minimum thickness equal to the probe spacing. The smallest spacing commercially available is of the order of 12 cm to 16 cm. This is commonly employed for short spaced density tools. In analogue systems, the mechanical limitations of data recording can dramatically increase the minimum resolution. For example, a density probe with 16 cm spacing run at a typical logging speed of 6 metres per minute and a time constant of I se~ond will not register a full peak response with analogue recording for beds less than 46 cm thick. Digital systems have inherently greater bed resolution capabilities and allow the use of deconvolution operators to isolate the response from the coal bed itself. However, to utilise this advantage fully, a far greater knowledge of the response characteristics of the various tools is required.

Chemical Factors Affecting Accuracy The physical factors described previously may be overcome by improved logging and processing techniques. However, chemical factors, such as the coal's inherent moisture and maceral composition, also influence ash determinations from the density log and these cannot be resolved satisfactorily at present. (a) Moisture content normallr averaging _about S% (air dried) may vary slightly, causing variations in the bulk density of the coal. However, these variations are very minor and are negl1giole when comparing the statistical variation in normal probe response. (b) Coal generally contains varying proportions of different macerals, each with a different density. Two samples with identical ash content may have different density log responses, suggesting that the macerals in the coal may account for part of the difference. Figure 4 illustrates the different densities of coal macerals. Obviously, if there are abnormal quantities of fusinite or exinite at the expense of vitrinite, the density log and consequently, the simulated ash estimates, will be affected.

0°

e 0

Vrtmtes Exirites

□

Micri'ites

X FusiniteS 2

density gm.Jee

Figure 3

Density tool response as a function of relative density ( after Sam worth (1974))

Figure 4

Coal maceral densities

Relative Effects of Chemical and Physical Factors on Ash Estimation The relative degree of influence which the physical and chemical factors have on ash estimates can be seen by separating the direct raw density counts to ash conversion process into two steps: (a) Relative density against raw coal ash, and (b) raw density tool count rate against relative density. The influence of inherent moisture and differing maceral densities is apparently responsible for the scatter in the ash against relative density cross plot (Figure 5). As illustrated, ash and relative density have a very close relationship, indicating that chemical factors do not introduce the significant errors seen in the density counts against ash cross plots. However, there is significant scatter in the raw density counts against relative density graph (Figure 6). This suggests physical factors are the major cause of inaccuracies in the conversion of density counts to raw ash content.

.... . ...... "' "' .,.s,," •••

Figure 5

Ash versus RD crossplot (Felton area near Toowoomba)

Figure 6

,..

JOO

long spaced densUv coun1 rote (cps)

Raw density count rate versus relative density ( Felton area near Toowoomba)

Volatile Matter Estimates of volatile matter content for raw coal were attenpted using neutron log data. The neutron count rates were initially corrected for borehole diameter by the application of a rudimentary, practically determined relationship. A linear n lationship was established between volatile m 1tter content and corrected neutron count rate. Accuracy of the resulting estimates was approximately ± 10% for all coal beds in a given prospect. Estimates are typically accurate to ± 5% for any one coal bed. Moisture Content o accurate log calibrations could be achieved for the inherent moisture content of raw coal. Because o f the relatively high neutron moderation of the cur :ent neutron-neutron logging method in water-rich seam intervals, the probe sensitivity to fairly low inherent coal moisture (less than 5%) is ver:r poor, thus making accurate log calibratio11s extremely difficult. Reeves (1!176) suggests the development of a slimline dncentralised epithermal neutron tool similar to those currently used in the petroleum industry vrould give better definition of moisture content within strong neutron absorbers. This developmimt would only be necessary in an area whern moisture content is an important parameter in the economic appraisal of a coal deposit. Specific Energy Geophysically derived estimates of specific energy ha•re been calculated from the density log via thE intermediary of the ash determination. Because o f the method of estimation, the degree of accuracy is of the same order of magnitude as for the ash estimates. Cumulath·e Float Fraction Yield As with specific energy, estimates of cumulative float fraction yield have been obtained from the geophysicilly derived ash determinations. These estimates ue the least accurate of the coal quality paramete1s determined, with absolute accuracies typically 1.0 ± 15%. The greatest obstacle in obtaining meaningful yield estimates is that no single relationship between ash content and clean coal yield can be applied to an exploration area as a wholf, since there is a unique relationship between t h.e two parameters for each coal bed. One of th,i components of the unique ash/yield relationship is the maceral content of different coal plies. As the maceral content of coal is independt nt of the ash content and macerals have different densities, coal samples of equivalent ash ,:ontent will not necessarily have similar float fraction yields when washed. Hence, coal plies from one borehole may have similar ash estimates, but quite different float fraction

yields. Errors in calculated estimates may also arise due to calibration of the original log with laboratory results. When washing the coal at a given relative density in the laboratory, it is possible that "dense" coal macerals or "stone" fragments that would usually sink, may float due to the collective buoyancy of other coal macerals attached to them.

Presentation of Quality Parameter Estimates Estimates of ash content, relative density, float yield and specific energy are presented in the form of a "pseudo-washability table". An example of such a table for the high ash coal prospect area. accompanied by the respective seam ash profile, is shown in Figure 7. Improvements in logging tool and proc·essing technology are needed to increase the accuracy of these estimates to a stage where they become an acceptable source of exploration data.

FUTURE DEVELOPMENTS IN LOGGING TECHNOLOGY hnprovements to Existing Tools Experience with the application of downhole geophysical logging in coal exploration has revealed several areas in which developments or improvements are necessary. The following suggestions include developments that are possible with present technology. (a) New density tools with source-detector spacings of the order of 5 cm or less should be developed to improve the effective

(b)

(c)

(d)

(e)

resolution. This development must be accompanied by the application of gamma ray spectral analysis technology to the logging probes. Such technology is necessary with small spacings to compensate for the effect of borehole cavities. An accurate caliper log run with, and on the same side of the borehole as, the density log. This would enable an accurate correction to be made to the density log for variations in borehole size. A decentralised neutron tool should be implemented to reduce the large effect that borehole fluids have on the log, especially in cavities. An accurate caliper log run with this tool would again be very useful in obtaining "true" log responses. In complex structural regions, a dipmeter log showing formation dips would be very useful. Dipmeter tools are already an integral part of large diameter logging services. Such a log would prove useful in determining the type and pattern of sedimentation, crossbedding and, in some circumstances, fault locations within the borehole. With the increasing importance of groundwater hydrology and environmental control, further research on the use of temperature and flow meter logs should be started. In particular, slimline flowmeter tools should be constructed and developed in an attempt to quantify water movements in aquifer zones. Such information would also be of use in mine planning.

tllLLnl:RRAN JOINT VEN'IURE ~ORCHOLE NO. FED 018 AS!' DETEP,'IN>TION FRO!•• LONG SPACED DENS I TY LOG LOG SCALC l: 50 ASH ?LO'! SCALE lOt/c,r . BASELI!lE VALUE C\ 20

19 ~lORKIN<' SECTION FLOAT FRACTION 20

l . 60 l. 80 2 . 00 RAW

16 . 0 .. .

20 . 9m.

YIELD

ASH(\)

58 . 62 79 . 9 1 91 . 57 100 . 00

22 . 93 35 . 24 41 . 98 45. 56

s . G. (g/cc)

Figure 7

Typical computer printout assodated with on-site processing

l.45 l . 64 l .7 5 l.81

S . E. (l'J/kg) 23 . 73 19. 05 16. 4 9 15 .1 3

Introduction of New Tools There are several logging tools which at present are not in common use or are unavailable in Australia. The following describes some of the potential uses of these tools. (a) Focused Resistivity Log The focused resistivity log can resolve lithological boundaries to within 2 cm. It is consequently very useful in delineating bands within coal seams and determining intervals where core loss may have occurred during drilling. Zones of heat affected coal are also more easily identified using a focused rather than a single point resistivity log. Since it may delineate small scale features in coal beds, the focused resistivity log would be very useful for correlation. As well as high resolution , this log provides a measure of the coal's true resistivity which may be employed for coal quality estimation. (b) Sonic Log The sonic log can provide useful information for engineering, studies of underground and open cut mining operations. Combined with density information, data recorded from the sonic log may be used to compute compressional and shear wave velocities as well as shear, bulk and Young's moduli and Poisson's ratio. The sonic log is an excellent coal delineator and consequently is very useful for correlation. eutron Activation Logging (c) The use of neutron activation techniques has been reported by Caldwell (1969), Keys and Boulogne (1969), Senftle and others (1970), Wickmann (1971), Schultz and Smith (1973) and Senftle and others (1977) with varying success to determine elemental iron, aluminium, silicon, oxygen, copper and nickel quantities and will become important in coal logging applications if advances in technology make it possible to detect carbon, hydrogen, sodium and sulphur. If these advances are made, neutron activation log data may be useful in reducing the costs involved in conducting laboratory ultimate analyses on coal samples (although it should be stressed that it will never totally replace such analyses). Even at the present stage of development, the neutron activation technique provides a method of calculating the in situ ash content of coal seams. The proportion of iron, aluminium and silicon in a coal ply normally constitutes approximately 95% of the ash content of that ply, and, by determining the quantities of these elements, an estimate of ash content may be made.

ACKNOWLEDGEMENT The authon would like to thank the Coal Division of AMAX Iron Ore Corporation for permission to present t 1eir findings regarding the application of digital bc,rehole logging to coal exploration.

REFERENCES Caldwell, R.L., 1969. uclear logging methods, Isotopes and Radiation Technology, VoL 6, No. 3 . p 257 - 273. Keys, W.S. & Boulogne, A.R., 1969. Well logging with c:alifornium - 252. SPWLA I 0th Annu1l Logging Symposium, May 25 - 28. Kowalski, J .J. & Holter, M.E., 1976. Coal analysis from well logs, C/M Bulletin, June C976, p 99 - 103. Lavers, B.A. & Smits, L.J.M., 1976. Recent developments in coal petrophysics, SPWLA 17th Annual Logging Symposium, Denver, Colorado, USA,June 9 - 12, 1976. Peeters, M. & Kempton, .H., 1976. Wireline loggir.g for coal exploration in Australia, 25th .rnternational Geological Congress, Sydnc:y, August 1976. Reeves, D.F:., 1971. In situ analysis of coal by borehole logging techniques, CIM Transactio11s, VoL 74, p 61 - 69. Reeves, D.F'.., 1976. Development of slimline loggir.g systems for coal and mineral exploration, SPWLA 17th Annual Logging Symposium, Denver, Colorado, USA, June 9-12. Samworth, J.R., 1974. The radiation density log applied to the resolution of thin beds in coal measures, 3rd European Symposium SPWLA, London Chapter, October 14 - 15. Schultz, W.A.. & Smith, H.D., 1973. Laboratory and field evaluation of a carbon/oxygen (C/O; well logging system, SPE - A/ME, 48th Annual Fall Meeting, Las Vegas, Nevada, 30 September - 3 October. Senftle, F .I .. and others, 1970. Use of Cf2 5 2 for mine1al exploration: comparison with acceli:rators for in situ neutron activation of sil1·er, Isotopes and Radiation Tech. Vol. :r, No. 4.

59 Senftle, F.E. and others, 1977. Importance of neutron energy distribution in borehole activation analysis in relatively dry, lowporosity rocks, Geoexploration No. 15. Wichrnann,P.A., 1971. Neutron activation for elemental determination in boreholes, SPWLA 12th Annual Logging Symposium.