The sampling of coal can be a difficult task in that coal is a heterogeneous material. Samples are the representative fractions of a body of material that are acquired for testing and analysis in order to assess the nature and composition of the parent body. They are collected by approved methods and protected from contamination and chemical change. Such samples should be differentiated from those materials collected in ways that may not be truly representative of the coal from which they have been collected. These materials may still be useful but should be regarded as specimens rather than samples. Coal samples may be required as part of a greenfield exploration programme to determine whether the coal is suitable for further investigation, or as part of a mine development programme, or as routine samples in opencast and underground mines to ensure that the quality of the coal to be mined will provide the specified run of mine product.
In situ coal samples are taken from surface exposures, exposed coal seams in opencast and underground workings, and from drill cores and cuttings.
Ex situ samples are taken from run of mine coal streams, coal transport containers and coal stockpiles.
Sampling may have to be undertaken in widely differing conditions, particularly those of climate and topography. It is essential that the sample taken is truly representative as it will provide the basic quality data on which decisions to carry out further investigation, development, or to make changes to the mine output will be made. It is important to avoid weathered coal sections, coals contaminated by extraneous clay or other such materials, coals containing a bias of mineralization, and coals in close contact with major faults and igneous intrusions.
In-situ sampling
Several types of in situ samples can be taken, dependent upon the analysis required
Grab sampling
Generally this is a most unsatisfactory method of obtaining coal for analysis, as there are no controls on whether the coal is representative, and can easily lead to a bias in selection, for example the bright coal sections attract attention. However, grab samples can be used to determine vitrinite reflectance measurements, as an indicator of coal rank.
Channel sample
Channel samples are representative of the coal from which they are taken. If the coal to be sampled is a surface exposure, the outcrop must be cleaned and cut back to expose as fresh a section as possible. Ideally the full seam section should be exposed, but in the case of thick coals (especially in stream sections), it may be possible to see only sections of the roof and coal immediately below, or the floor and coal immediately above. To obtain a full seam section under these circumstances, two or more overlapping channels will need to be cut, and the overlap carefully recorded. The resultant samples will consist of broken coal and will not preserve the lithological sequence. In opencast workings, the complete seam section should be exposed, and is less likely to be weathered than natural surface exposures. In underground workings, the seam will be unweathered, but the whole seam section may not always be seen, due to the workings only exposing the selected mining section of the seam.
To carry out a channel sample, the coal is normally sampled perpendicular to the bedding. A channel of uniform cross-section is cut manually into the coal seam, and all the coal within the cut section is collected on a plastic sheet placed at the base of the channel. Most channels are around 1.0m across and samples should not be less than 15kg per meter of coal thickness. Such channel samples will provide a composite quality analysis for the seam, that is an analysis of all the coal and mineral matter present in the seam as a whole. Although this is suitable for general seam quality assessment, more detailed analysis of the seam from top to bottom may be required. To achieve this, a channel ply sample is taken, which entails a similar procedure as for the whole seam channel sample except that the seam is divided into plies or subsections. Coal seams are rarely homogeneous throughout their thickness, most are divisable into distinct lithological sections. Plies are lithological subdivisions of the seam, each of which has a uniform character. When the lithology changes, such as at a clay parting in the seam, a separate ply is designated. Where the roof and floor of a seam are exposed, ply samples of at least 0.25m of roof material immediately above the seam and 0.25m of floor underlying the seam should be included in the samples. This will allow the effects of dilution on coal quality to be assessed. In general the thickness of coal plies should be a minimum of 0.1m and a maximum of 1.0m. In the case of banded coals containing alternating thin (<0.1m) layers of bright coal/dull coal/clay, the seams may be sampled as a series of composite plies, with the details of the individual layers shown on the record sheet. An interbedded non-coal ply greater than 0.25m in thickness may be regarded as a seam split and recorded as such. Ply samples should be at least 2.0kg where possible, it may be that the sample will be split into two fractions and one stored for later use. Once the outcrop or face is cleaned, a shallow box-cut is made for the total thickness of the exposed coal seam. Once this is completed, the seam is divided into plies, each of which is measured and recorded on are cord sheet. The channel sample record sheet should show the following information.
- Record card number.
- Map or aerial photograph number on which locality is located.
- Location of sample point, grid reference or reference number.
- Description of the locality, stream section, working face, etc., including dip, strike, coal seam roof and floor contacts.
- Extent of weathering, fracturing, mineralization, etc.
- Lithological description of each ply interval.
- Thickness of each ply interval.
- Designated sample number of each ply interval.
Space canal so be allocated on there cord card for analytical details, that is proximate analysis, to be added later to complete the record. The fresh surface is then sampled as a channel cut from top to bottom, cutting and collecting all material from each ply section in turn. Each ply sample should be sealed in a strong plastic bag immediately after collection to prevent moisture loss and oxidation. All sample bags must be clearly labelled with a designated number, a copy of which should be placed in a small plastic bag inside the sample bag, and another attached to the outside of the sample bag. This number must be recorded on the channel sample record sheet. Because this task is invariably a dirty one, labels get wet, blackened and unreadable very easily, so it is essential therefore that care must be taken to ensure that the sample numbers do not get lost or obliterated during transit to the laboratory, as unidentifiable samples are useless and an expensive waste of time. The advantage of channel ply sampling is that not only can the analysis of the individual plies be obtained, but also by combining a fraction of each ply sample, a whole seam composite analysis can be made. An example of a channel ply sample from a surface exposure is illustrated in Figure 5.3, which shows a channel cut to expose fresh coal, and then a thinner channel (0.25m wide) cut from the fresh coal from which ply samples are collected for analysis.
Pillar samples
In underground coal mining, samples of large blocks of undisturbed coal are taken to provide technical information on the strength and quality of the coal. These pillar samples are taken when a specific problem may have arisen or is anticipated. Such samples are taken in much the same way as whole-seam channel samples except that extra care is required not to disturb the cut-out section of coal during removal. Samples are then boxed and taken to the laboratory. Pillar sampling is a long and arduous business and is undertaken only in special circumstances, such as when mining becomes difficult or new roadways or faces are planned.
Core samples
Core sampling is an integral part of coal exploration and mine development. It has the advantage of producing non-weathered coal including the coal seam floor and roof, and unlike channel samples, core samples preserve the lithological sequence within the coal seam. First, the borehole core has to be cleaned if drilling fluids have been used, and then lithologically logged. Following this, the lithological log should be compared with the geophysical log of the borehole to select ply intervals and to check for core losses and any other length discrepancies. Once the core has been reconciled to the geophysical logs and the ply intervals have been selected, sampling can commence. Core ply samples are taken in the same ways for surface channel ply samples, again a ply sample of the coal seam roof and floor (up to 0.25m) is taken to determine dilution effects. Then the individual plies are sampled, making sure no core is discarded. As in the case of surface samples, bright coal tends to fragment and make up the finer particles that may easily be left in the core tray. The samples are bagged and labelled as for surface ply samples, and the sample numbers recorded on the core-logging sheet. Large diameter cores may be split lengthways with a bolster chisel and then one of the halves ply sampled, the other being retained for future analysis.
Cuttings samples
This method of sampling is considerably less accurate than that of core sampling. As with core samples, cuttings are unweathered and are a useful indicator as to the general nature of the seam. Air flush and mud flush noncore drilling is a quicker operation than core drilling and will produce cuttings for each horizon encountered in drilling. In the case of mud flush cuttings they will need to be washed to remove any drilling fluid before sampling. Cuttings are usually produced for every metre drilled, those cuttings returns that are all coal may be collected, bagged and numbered in the same way as channel samples. The depth to the top and bottom of the seam sampled should be determined from the geophysical log. The drawback with using cuttings samples is that only a general analysis of the seam can be made, and even this is unlikely to be truly representative. Contamination from strata above the coal also may be included, and a close study of the geology will determine whether this is so.
Specimen samples
Orientated specimens of coal may be collected so that their precise orientation can be re-created in the laboratory. The dip and strike of the coal is marked on the specimen before removal. This method is commonly used for studies of the optical fabric of the coal, or of the structural features in the coal.
Bulk samples
Bulk samples are taken from outcrops, small pits or minishafts (i.e. 2m diameter shaft excavations). A bulk sample is normally 5–25t and is taken as a whole seam channel sample on a large scale. Such a bulk sample is taken in order to carry out test work on a larger scale, which is designed to indicate the coal's likely performance under actual conditions of usage. Steam coal samples are taken for small combustion tests in a pulverized fuel (pf) rig, to simulate conditions in a PF boiler. Pulverized coal firing is the combustion of powdered coal suspended as a cloud of small particles in the combustion air. Substantially more heat is released per unit volume in PF boilers than in stoker type boilers. Coking coal samples are taken to carry out moving wall oven tests, that is to determine how much the coal swells when it is combusted, thus putting pressure on the oven walls, which are constructed of uncemented brickwork. High-pressure coals are undesirable, and are normally blended with low-pressure coals to reduce the problem. In the United States, low-volatile coking coals (volatile matter (VM)=20–25%,SI=9) are high-pressure coals, whereas in general, high-volatile coals do not have such high pressures. It is significant that Gondwana coking coals are low-pressure coals, an important factor in Australia being able to export coking coals. Bulk samples are collected from a site already channel sampled, loaded into drums, numbered and shipped to the selected test centre.
Sample storage
In the majority of cases, the channel and core samples will be required immediately for laboratory analysis. However, there are circumstances where duplicate coal samples for future reference are taken. Usually the channel plies are divided into two or the cores are split and one half retained. If the duplicate samples are to be put into storage, this presents a problem because the exposure of the coal to air will allow oxidation to take place during storage and this will result in anomalous quality results when analysed at a later date. The usual procedure to prevent oxidation of samples is to store them under nitrogen or in water. To store in nitrogen, place a tube connected to a pressurized cylinder containing nitrogen in a plastic sample bag, then add the coal sample, flush the sample with nitrogen regulating the flow by means of a flow meter. The nitrogen has to fill the spaces between the coal fragments, so flushing with nitrogen is required for several minutes. One difficulty with this method is that nitrogen is lighter than air so inevitably some is lost in the process. Once the bag has been thoroughly flushed, it should be heat-sealed; no other form of sealing is anywhere near as effective. The coal samples can be as received or air dried and can be in the form of lump or crushed coal. It should be noted that for all in situ and ex situ samples, the top size to which any sample is crushed to is important in determining the weight of the sample required. The size of the sample is calculated as: 5.24×mean particle size=xkg (where mean particle size is top size×bottom size). 5.24 is an empirically determined number quoted in BS1017-1 and Australian Standard 4264 (Appendix 1). A cheaper method of storage is by immersing the channel or core sample in the form of lump coal in water. This method has the advantage over storing in nitrogen in that it preserves fluidity of the coal, but it does present handling problems when the sample is required. The sample will have to be air dried before analysis can begin. Samplescanbekeptbythesemethodsfor1–2yrbefore analysis.
Ex-situ sampling
The object of collecting coal samples after mining is to determine the quality of coal actually being produced. This coal may differ significantly from the in situ seam analysis in that not all of the seam may be included in the mining section, or that more than one seam may be worked and fed to the mine mouth and mixed with coal from other seams. In addition there may be dilution from seam roof and/or floor contamination that becomes part of the mined coal product. The mined coal is broken up and therefore contains fragments that vary a great deal in size and shape. Representative samples are collected by taking a definite number of portions, known as increments, distributed throughout the total quantity of coal being sampled. Such increments represent a sample or portion of coal obtained by using a specified sampling procedure, either manually or using some sampling apparatus. The various practices used in collecting ex situ samples and the mathematical analysis of the representativeness of samples, i.e. quality control. Increments are taken using three methods.
1. Systematic sampling, where increments are spaced evenly in time or in position over the unit.
2. Random sampling, where increments are spaced at random but a prerequisite number are taken.
3. Stratified random sampling, where the unit is divided by time or quantity into a number of equal strata and one or more increments are taken at random from each.
It is good practice that whatever the method used, duplicate sampling should be employed to verify that the required precision has been attained. Ex situ coal sampling is carried out on moving streams of coal, from rail wagons, trucks, barges, grabs or conveyors unloading ships, from the holds of ships and from coal stockpiles.
Hand sampling from streams is carried out using ladles or scoops, the width of the sampler should be 2.5 times the size of the largest lump likely to be encountered; however, this type of sampling is not suitable for coal larger than 80mm. For larger samples mechanical sampling equipment is used, where moving streams of coal (conveyors) are sampled by: (a) falling stream samplers, which make either a line a traverse across the coal convey or in a straight line path perpendicular to the direction of flow, or opposite to the direction of flow, or in the same direction of flow, or they make a rotational traverse by moving in an arc such that the entire stream is within the radius of the arc; (b) cross-belt samplers, which move across the belt pushing a section of coal to the side while the belt runs; (c) the stop belt method, whereby the conveyor is stopped and all coal occurring within a selected interval, usually a couple of metres, is collected. The correct increment selection occurs when all the elements of the transversal cross-section are intercepted by the sampling cutter during the same length of time. This should avoid any increase in error. These sampling systems are checked for bias by using a reference sampling method as recommended by BSISO 13909 or ASTM D2234. Wagons and trucks are sampled by taking samples from their tops by means of probes, or by sampling from bottom or side door wagons during discharge,or sampling from the exposed face of coal as the wagons or trucks are tipping into bunkers or ships, or wagons being emptied via tipplers. Ships are sampled either from conveyors loading and unloading coal, at a point where bias can be avoided, or from the hold of the ship. Samples from the hold, are taken every 4m of the depth of the coal within the hold.Itisimportanttoestimatetheproportionoffine and lump coal in the consignment. It should be noted that free moisture, if present, will tend to settle towards the bottom of the hold. This increase of moisture with depth makes it difficult to collect samples for moisture content determination. 4. Sampling from barges is the same as for ships except that if the depth of coal is less than 4m it should be sampled onboard during unloading, once the bottom of the barge is partially uncovered. 5. When the preferred procedure of sampling from a conveyor belt during stocking and unstocking cannot be used, then the stockpile is sampled based on collecting increments spaced as evenly as possible over the surface and layers of the stockpile. Sampling is by means of probes or by digging holes. If the stockpile is known to consist of different coals piled in separate areas of the total pile, a separate gross sample must be taken from each such area. The stockpile should be divided into a number of portions, each 1000t or less from which a separate sample with a specified number of increments is taken. This normally takes along time to accomplish, but can be speeded up if automated auger units are employed. It is important that all levels in the stockpile are sampled.