DETERMINING GAS PRODUCTION CHARACTERISTICS OF COAL SEAMS
IAN GRAY
Unwinding the mysteries of a coal seam methane reservoir can be substantially achieved without recourse to expensive production trials. This paper examines the options for obtaining the production characteristics of coal seams more efficiently in terms of cost and time, than production tests. The techniques reviewed include DST testing from within HQ wireline core rods, sorption pressure testing, measurement of diffusion coefficients and core permeability testing.
The information gained from the suite of tests includes diffusion coefficient, coal block size, matrix permeability, fracture permeability, reservoir pressure, sorption pressure, gas content and the effects of shrinkage on permeability.
THE CHARACTERISTICS OF COAL SEAM RESERVOIRS
Coal Seams are very complex gas reservoirs. The gas they store is held unconventionally and they may show several levels of permeability. Sometimes permeability increases with extraction whilst in other cases it decreases. Some reservoirs may be treated as homogeneous and modelled using derivatives of conventional reservoir models. Others however behave as a series of domains of different characteristics separated by preferential flow paths. They may be recharged or drained via these flow paths. A knowledge of the reservoir characteristics is invaluable prior to starting gas production, or planning to de-gas a mine.
Gas Storage
The gas contained in coal seams is mostly stored in the coal itself by a process called sorption. Much less gas is stored in pore space either as free gas or in solution in seam water.
The reservoirs may be either saturated with water or contain free gas prior to extraction. The gas is stored in the coal at sorption pressure which is the equivalent of the bubble point in a conventional reservoir. The sorption pressure may be less than the reservoir pressure or in the case of a reservoir with free gas, at reservoir fluid pressure.
The gas storage characteristics are commonly represented by the volume stored (at STP) versus pressure relationship for a given temperature. Such relationships are often described by the langmuir isotherm which fits most but notably not all experimental sorption data.
The sorption isotherm is affected by gas type, and the history of how the gas came to be present in the coal. Cases have been noted in high rank coals where CO2 has been introduced of the CO2 being released first despite its apparently greater affinity for the coal. It is suspected that this is due to the CH4 being sited preferentially closer to the surface sites in the coal than the CO2 which is usually added during a cycle of igneous intrusion. Thus the history of CO2 then CH4 generation and CO2 intrusion (probably as hydrothermal events) becomes important from a time perspective in determining the nature of the sorption isotherm and the gas type released.
The sorption isotherms measured in the laboratory are known to be affected by their moisture content (saturation).
This is almost certainly related to the coal being either gas wet or water wet and is associated with competition for sites at molecular level.
The consequence of this is that simple laboratory isotherm testing is not adequate to determine the gas storage characteristics in a seam.
Thus measuring a gas content by core desorption and then re-testing the ground and dried core for a sorption isotherm and using the combined information to arrive at a sorption pressure may be significantly in error. This error is not solely due to the problems in measuring the sorption isotherm but is also brought about by the uncertainties in the measurement of gas content. Specifically the estimation of the gas lost whilst pulling the core remains an issue. Together these uncertainties may lead to significant errors in estimates of sorption pressures especially in the areas of higher gas content where the sorption isotherm shows a high rate of change of pressure for a given change in stored volume.
In essence the requirement is therefore to measure the original gas content and pressure at a few points and thus derive a native sorption isotherm.
Gas Movement
Diffusion
Gas movement appears to occur from within the coal to the cleat structure by a process of diffusion. The speed of the diffusion is dependent on the diffusion coefficient, the cleat spacing and the concentration gradient.
It is relatively straightforward to determine the diffusion coefficients by examining the rates of gas release from core and by assuming the core is a uniform cylinder.
Typically the core displays two diffusion coefficients, an initial one and then a long term one. Typically the short term diffusion coefficient is an order of magnitude greater than the long term coefficient. Occasionally a third diffusion coefficient appears to exist during the long term diffusion process.
The change between short and long term diffusion coefficient is quite marked. It is not yet known whether this change is brought about by pressure changes and as such is essentially a pressure sensitive effect on a quasi-darcy flow or is truly a separate gas release process. The assumption that the coal core is a cylinder must be considered inaccurate in the case of cleated coal.
For these reasons the determination of diffusion coefficients are the subject of ongoing research.
As the rate of diffusion is strongly related to the size and shape of the coal blocks in situ from which the gas is escaping the determination of these dimensions is important.
Matrix Permeability
Matrix permeability exists in the cleats where two phase (gas and water) darcy flow takes place down a potential (principally pressure) gradient. During the drainage cycle the permeability changes due to changes in water saturation and due to effective stress variations in the coal. Effective stress is the difference between the total stress, which varies with direction, and the fluid pressure.
Permeability reduces with effective stress in coals. The softer and more cleated the coal, the more dramatic this reduction is. Thus as a coal seam gives up fluid the effective stress might be expected to increase due to reduced pore pressure.
This is often not the case.
The reason for this is that the effective stress is not solely related to the fluid pressure but also to the gas content of the coal. It has been shown repeatedly that the change in linear dimension of a piece of coal is dependent on changes in gas content. With reference to seam drainage this dimensional change is one of shrinkage.
Most coal seams are bounded by significantly stiffer roof and floor rocks and therefore the the lateral dimension of the seam is more or less constant but the vertical dimension is free to move. In this environment the total vertical stress in the seam can be expected to remain constant but the effective vertical stress will increase with fluid removal.
In the absence of shrinkage lateral effective stress could be expected to increase due to a reduction in seam fluid pressure. The lateral effective stress increase would be less than the vertical stress increase because the seams are laterally extensive and usually bounded by much stiffer roof and floor rocks. Shrinkage tends however to reduce the coal dimension and with it the effective stress.
The two effects on effective stress are opposing and without adequate testing it is not possible to determine whether matrix permeability will increase or decrease. From field experience both cases are known to exist. The matrix permeability can in some reservoirs change by orders of magnitude with gas production. This change may increase or decrease permeability. Knowing which is vital. Because cleats are directional and stress is directional the matrix permeability is also directional, though fracture permeability may mask this feature in many instances.
Fracture Permeability
Frequently major joint sets also exist within the coal seam and lead to a second level of greater permeability. Major faults may also transect the coal seam acting as barriers, fluid sources or sinks.
Domains
Some coal seam reservoirs behave as though they contain quite different domains with significantly differing properties. These domains are known in some instances to be bounded by faults and are thus structurally different.
AN APPROACH TO RESERVOIR ASSESSMENT
The author’s company has developed a procedure for determining coal seam reservoir behaviour. This in essence involves the following process.
Initially a general geological appraisal of an area is made to determine the factors that may influence gas storage or movement. Once the basic geology is understood the programme focuses on drilling and downhole testing.
Use is made of slimhole drilling with wireline coring
(usually using the Longyear HQ-3 system) through the coal bearing strata. Slim holes are cheap to drill and wireline core retrieval provides core quickly for gas content measurement, the assessment of diffusion coefficient, shrinkage and permeability testing. Coal thickness can be readily determined and a hole is available for stress and in situ permeability testing.
Stress Measurement
Stress measurements are made in the non-coal strata utilizing overcoring or hydrofracture. Measurements are usually made in rock because reliable stress measurement is difficult to achieve in coal by either method. The results are then interpreted in terms of tectonic strain through the sequence so that the coal stress may be estimated on the basis of its modulus and Poisson’s ratio.
Core Desorption
Core is withdrawn from the core barrel by wireline and placed in a canister for initial desorption. The initial desorption process is required so that an estimate may be made of the lost gas on core retrieval. The initial period of desorption (~30 minutes) also provides information on the initial diffusion coefficient.
The core should then be transferred to a water-filled pressure vessel and sealed and allowed to come to equilibrium pressure. Desorption may then be continued so that the remaining volume of gas contained in the core can be measured and to permit the estimation of the long term diffusion coefficient. It is generally undesirable to grind the core to get a residual gas content as this destroys the core.
Establishing a curve fit based on a long term diffusion coefficient enables the remaining gas to be estimated with good accuracy.
The retained core is kept for examination for cleats, permeability testing and shrinkage testing. Some may also be used for petrological determination.
Drill Stem Testing for Matrix Permeability, Reservoir Pressure, and Sorption Pressure
The next preferred operation is to undertake a DST test using a piece of equipment specifically developed for coal seam methane operations. The tool’s operation is shown in Figures 1 to 8. This tool’s operation may be summarized as follows:
1) Core is broken off and the inner barrel is retrieved.
2) The rods are withdrawn above the seam to be tested.
3) The DST tool is pumped down inside the rods to seat on the landing ring. In this state one packer extends through the core bit whilst the other remains inside the core barrel. As the drill string does not need to be pulled or another string run with the tool a significant time saving is achieved over other DST tools.
4) Pump pressure is raised to inflate the packers. This mode of operation makes the tool independent of electrical cables or and inflation line, considerable simplification. It also permits its use at significant depth.
Two options now exist, either to inject fluid into the seam or alternatively to produce from the seam.
In the case of injection:
5a) The drill string is rotated a ¼ turn to open the bottom valve.
6a) Water is injected at a constant pressure whilst flow is monitored. Several pressure steps are used.
In the case of production:
5b) The string is lowered to open the dump valve.
6b) Compressed air is applied to the top of the drill string to displace drilling fluid out of the dump valve.
7) The string is raised to close the dump valve and compressed air pressure is released.
8) A gas flowmeter is connected to the top of the string.
9) The string is rotated to open the bottom valve.
10) Fluid flow takes place from the seam and into the drill string. Gas flow is monitored by the gas flow meter whilst liquid is monitored by pressure build up in the string.
During the entire process pressures are automatically monitored in the bottom hole zone, in the annulus, in the string and in the packers.
11) The string is rotated to close the bottom valve.
12) A pressure build up takes place.
13) The string is rotated to open the bottom valve.
14) Pressures are permitted to equalize.
15) A wireline overshot is lowered into the string and latches onto the spearhead. It is pulled and opens a valve which deflates the packers.
16) The entire assembly is hoisted out of the drill rods on the wireline.
17) The chronological record of bottom hole pressure, annular pressure, packer pressure and pressure within the drill string is retrieved electronically from the assembly for analysis.
The injection test is useful in a water saturated seam because it permits the single phase permeability to be established.
The step varied injection pressures also permit an apparent skin factor to be determined at various pressures. In our experience the skin factor in coal is pressure dependent and can be viewed as being associated with varying effective stress around the wellbore. Indeed by making assumptions about the pressure distribution and shape of the sorption pressure effective stress relationship (linear plot of log k vs effective stress) it is possible to derive a permeability pressure relationship from injection testing a well.
A production test is useful because it permits production of seam fluids and can permit the assessment of sorption pressure and reservoir pressure. In the event that gas is produced then the analysis for permeability is complicated by desorption effects and re-absorption of gas into the coal during the pressure build up phase. The process is one of gas being produced from the coal near the wellbore during the production phase. At the same time water and gas are produced along the cleats. When the well is closed pressures rise and the gas is re-absorbed back into the coal. When re-absorbtion is complete the area around the wellbore re-saturates and this is associated with a kick in the pressure response with time. This kick corresponds with the sorption pressure existing at that time.
Cleat Spacing
This can be assessed by visual examination of core or by optical/acoustic examination of the wellbore. In the event that cleats are not visible in sufficient number to determine their spacing and orientation then their effect on reservoir performance can only be assessed by a history matching of production and pressure with a simulator. This approach is fraught with problems of lack of uniqueness of solution.
Native Sorption Isotherms
The sorption pressure derived from the DST test and the sorption pressure gained by sealing the core in a water filled canister on surface when combined with a knowledge of gas contents at these two stages provides enough data to permit the form of a native sorption isotherm to be derived. It is still useful to undertake a laboratory isotherm test for comparison but preference should always be given to the native sorption isotherm especially in the case of mixed gases.
Fracture Permeability Assessment by Interference Tests
Once the DST test is complete it is usual practice to place packers fitted with pressure transducers over the seams in boreholes affected by subsequent testing or production. An alternative is to grout pressure transducers in place but this can lead to problems with channelling in the grout if the reservoir pressures are greater than hydrostatic. The placement of such pressure transducers is essential for establishing fracture permeability and is invaluable for reservoir monitoring during subsequent extraction.
The placement of pressure transducers in monitoring holes around an injection or production hole enables the testing to take place for fracture permeability. In a saturated system injection yields the best results because flow remains in a single phase and the permeability results are easy to analyse.
If free gas exists in the reservoir it is a better option to produce from the well. Such an interference test can be made part of a production operation.
Shrinkage/Stress Effects
The assessment of the effects of shrinkage and effective stress on permeability can be made by two laboratory/field techniques or by a production test. The latter takes a long time while the field/laboratory tests can be undertaken relatively quickly. Two effects need to be known. The first is the effective stress/permeability relation and the second is the gas content shrinkage behaviour. In addition a knowledge of the stress field is required.
The effective/stress permeability relation can either be obtained through careful analysis of the change in skin effect brought about by varying injection pressures to a well or by laboratory testing. Both are useful and may yield similar or varying results. The laboratory testing involves placing the vertical core in a triaxial test rig and passing gas transversely across the core. The total vertical stress is maintained whilst the lateral confining stress is varied. The use of helium or nitrogen as the gas restricts the effects of shrinkage/growth of the coal thus enabling the derivation of a straightforward effective lateral stress/permeability relation.
Coal expansion and shrinkage is best assessed by strain gauging core and by placing it in a vessel which can be filled with gas at a series of controlled pressures. Between changes in pressure the core is allowed to approach equilibrium and the strain changes are recorded. Equilibrium is never reached however and numerical assumptions on the nature of absorbtion need to be made to arrive at the final strain/gas pressure relationships. The original design of the core permeability test rig was to permit strain gauged core to be used in the cell and to fully gas the core and then to bring it up to in situ stress levels. The intention was then to hold a constant lateral strain whilst reducing gas pressure and permitting flow through the core, thus providing a measurement of changes in permeability with varying gas pressure and levels of sorption. This has never been able to be realized however because of the long duration required for the tests. To undertake such testing the rig could be easily tied up for 3–6 months on a single piece of core. It was nearly impossible to keep strain gauges operational under conditions of stress, gas and moisture for this period of time.
In addition it was grossly inefficient to tie up the permeability rig for this period hence the use of the separated effective stress/permeability and shrinkage/sorption tests.
The alternative of permeability testing at varying effective stresses combined with separate sorption/shrinkage testing has proven faster and more cost efficient.
It is worth noting that not all coals show significant changes in effective stress with permeability. Some of the harder coals show negligible change.
Domains
Some coal seam methane reservoirs are noted as having significant spatial changes in behaviour. These can commonly be related to changed stress situations or to the history of groundwater movement. Geological studies may reveal potential zonal differences. Often however these changes are found as the result of drilling and testing. Only when the tests have been performed can the geology be re-examined and conclusions drawn about which domains of varying reservoir behaviour may exist.
CONCLUSIONS
This paper presents the techniques required to establish the behaviour of coal seam methane reservoirs without recourse to production testing. Emphasis has been placed on measuring relevant parameters rather than attempting to history match production data utilizing a reservoir simulator.
The latter approach suffers from a lack of uniqueness of solution that may lead to significant errors in estimation of long term reservoir performance.
The test techniques described yield results on reservoir behaviour more quickly and cheaply than production testing.