David N. Dewhurst

Compaction‐driven evolution of porosity and permeability in natural mudstones: An experimental study

This paper describes a series of experiments designed to investigate the influence of lithology on the compactional loss of porosity and permeability in mudstones. Two intact samples of London Clay with clay fractions of 40% and 67% were compacted to 33 MPa effective stress. Clay fraction, permeability, porosity, pore size distribution, and specific surface area were measured and their evolution was monitored throughout the compaction process. Electron microscopy was combined with mercury porosimetry to trace the collapse of the pore structure with increasing effective stress. In both cases, porosity loss occurred primarily by the collapse of large pores. This process is more obvious in the coarser-grained sample because throughout the compaction process it has a much broader range of pore radii and a much greater mean pore radius. Consistent with the pore size distributions, the permeability of the coarser sample ranges from ∼ 10−10 m s−1 to 10−12 m s−1 while that of the finer-grained sample ranges from ∼4 × 10−12 m s−1 to 5 × 10−14 m s−1 during progressive compaction from 2 to 33 MPa. The compressibility of the finer-grained sample is greater than that of the coarser-grained sample (0.15 as opposed to 0.07). However, in both cases the compressibility is much lower than that inferred for lithologically similar samples compacted over geological timescales. The demonstration that both porosity and lithology (clay fraction) influence the permeability of mudstones should allow the development of more realistic porosity-permeability relationships which take into account lithological variations exhibited by mudstones.

Permeability and fluid flow in natural mudstones

Abstract Mudstone permeabilities vary by ten orders of magnitude and by three orders of magnitude at a single porosity. Much of the range at a given porosity can be explained by differences in grain size; at a given effective stress, coarser-grained mudstones are more permeable than finer-grained mudstones, although the difference diminishes with increased burial. Pore size distributions illustrate why more silt-rich mudstones are more permeable than finer mudstones and also show that the loss of porosity and permeability with increasing effective stress is driven primarily by the preferential collapse of large pores. Pore size distributions can also be used to estimate permeability rapidly. None of the existing models are ideal and need to be adjusted and validated through the acquisition of a much larger permeability database of well-characterized mudstones. We also examine the role of faults and fractures as fluid conduits in mudstones. The occurrence of microscopic hydrofractures is inferred from the observation that fluid pressures in sedimentary basins rarely exceed minimum leak-off pressures. The extent to which microfractures enhance mudstone permeability, both instantaneously and over longer periods of geological time, is poorly constrained. Although fault zones in mudstones have generally low permeability, there is abundant evidence for episodic flow along faults in tectonically active regions. The role of faults as fluid conduits during periods of tectonic quiescence is less certain, and the timing and extent of any enhanced permeability and enhanced flow are not well known. In general, conditions conducive to fluid flow along muddy faults include an increase in the activity of the fault, high fluid pressures within the fault zone and the extent of overconsolidation and lithification of the mudstones.

The development of polygonal fault systems by syneresis of colloidal sediments

Abstract Polygonal fault systems occur in numerous sedimentary basins worldwide, are generally located on passive margins in onlap fill units and tend to comprise the finest grained sediments in this geological setting. These fault systems have been most thoroughly described in the central North Sea basin and the detailed structure shows a significant correlation with lithological variations, both vertically and laterally. Extension measured in stacked decoupled tiers of polygonal faults correlates positively with both clay fraction and smectite content. Lateral facies variations are also observed and indicate that time-equivalent sequences upslope from the smectite-rich polygonally faulted sediments are coarse-grained, clay-poor and undeformed. This leads us to believe that the structure and geometry of the fault system are controlled by the colloidal nature of the sediments, and that the volumetric contraction measured on seismic sections can be accounted for by syneresis of colloidal smectitic gels during early compaction. Syneresis results from the spontaneous contraction of a sedimentary gel without evaporation of the constituent pore fluid. This process occurs due to the domination of interparticle attractive forces in marine clays, dependent on environment, and is governed by the change of gel permeability and viscosity with progressive compaction. The process of syneresis can account for a number of structural features observed within the fault systems, such as tiers of faults, the location of maximum fault throw and growth components at upper fault tips. As such, this paper represents the first attempt to correlate microscale properties of clay-rich sediments to their macroscale seismic character.

Influence of clay fraction on pore‐scale properties and hydraulic conductivity of experimentally compacted mudstones

We report the results of a series of hydraulic conductivity tests carried out on seven natural, well-characterised specimens of London Clay mudstone. The clay fractions of the samples range from 27% to 66% and enabled a test of the influence of clay fraction on the hydraulic conductivity, pore size distribution, compressibility and specific surface area of natural mudstones. Hydraulic conductivities were determined at effective stresses between 1.5 and 33 MPa. Hydraulic conductivities of clay-rich samples (49–66% clay fraction) decreased from ∼10−11 m s−1 to ∼10−14 m s−1 over a porosity range of 48% to 25%. At a given porosity the hydraulic conductivities of two silt-rich samples (27 and 33% clay fraction) were 40–250 times greater than those of the five clay-rich samples. Variations in hydraulic conductivity are directly related to pore size distributions and are accurately predicted by a model which uses pore size distribution as its primary input. Clay-rich samples have unimodal pore size distributions with modal throat radii around 60–120 nm. Silt-rich samples have bimodal pore throat size distributions. One modal size is similar to that observed in clay-rich samples with a second modal value at 3–6 μm. Compaction under effective stresses up to 10 MPa results in the preferential collapse of larger pores, so that the rate of loss of hydraulic conductivity is greater in the silt-rich samples. Differences in hydraulic conductivity between silt-rich and clay-rich mudstones therefore decline with decreasing porosity. The range of porosity-hydraulic conductivity relationships means that hydraulic conductivity is not easily predicted from porosity alone; additional constraining parameters such as grain and pore size distributions are required.

Microstructural and petrophysical characterization of Muderong Shale: application to top seal risking

Analysis of the Muderong Shale from the Carnarvon Basin suggests the shale is dominated by interstratified illite–smectite with a high percentage of illite interlayers. Capillary pressure measurements indicate that gas columns of c. 250 m could be sealed by such shale, although the choice of drying method used does influence the accuracy of this calculation. Freeze drying yielded the most consistent threshold pressure results, whereas air drying and vacuum drying showed a greater range of values. Similar calculations in regard to carbon dioxide sequestration indicate column heights of between 550 m and 750 m could be retained. Column height variation is primarily dependent on the contact angle of supercritical carbon dioxide with shale. Microstructurally, the shale is clay supported, exhibiting differential compaction of clays around more rigid grains and containing numerous high aspect ratio discontinuous fractures. These fractures do not affect the capillary properties of the shale, even when injection is fracture-parallel, suggesting they are unlikely to influence reservoir-scale fluid-flow properties. Comparison of the Muderong Shale laboratory data with hydrocarbon column heights from Carnarvon Basin discoveries indicate that top seal failure by capillary breakthrough is unlikely given the maximum lengths of hydrocarbon columns encountered to date. Potential for top seal failure is more likely to be influenced by formation integrity, pore pressure and in situ stress conditions.

Geomechanical and ultrasonic characterization of a Norwegian Sea shale

Anisotropy of velocity in shaly overburden is known to cause significant problems for geophysical interpretation, including depth conversion and fluid identification. In addition, mechanical and dynamic elastic shale behavior is not well understood because few tests have been performed on well-preserved samples. Multiple stage triaxial tests were performed upon horizontal core plugs of a shale from the Norwegian Sea with a view to evaluating rock strength and the evolution of ultrasonic response during rock deformation. In addition, standard rock physical properties were characterized as well as composition. The shale microfabric is seen to be strongly laminated, with alternating thick clay-rich laminae and thin silt-rich laminae. Occasional microfractures are also noted parallel to these laminations. The shale has low friction coefficient and cohesive strength, and shows anisotropy of these parameters when the maximum principal stress is oriented parallel to and at 45 � to the microfabric. The orientation of the maximum principal stress parallel to the intrinsic fabric and microcracks was seen to significantly impact on velocity normal to the fabric as stress parallel to the fabric increased. S-wave anisotropy was significantly affected by the increasing stress anisotropy. Stress orientation with respect to fabric orientation was therefore found to be an important control on the degree of anisotropy of dynamic elastic properties in this shale.

Evaluating hydrocarbon trap integrity during fault reactivation using geomechanical three-dimensional modeling: An example from the Timor Sea, Australia

Three-dimensional (3-D) coupled deformation and fluid-flow numerical modeling are used to simulate the response of a relatively complex set of trap-bounding faults to extensional reactivation and to investigate hydrocarbon preservation risk for structural traps in the offshore Bonaparte Basin (Laminaria High, the Timor Sea, Australian North West Shelf). The model results show that the distributions of shear strain and dilation as well as fluid flux are heterogeneous along fault planes inferring lateral variability of fault seal effectiveness. The distribution of high shear strain is seen as the main control on structural permeability and is primarily influenced by the structural architecture. Prereactivation fault size and distribution within the modeled fault population as well as fault corrugations driven by growth processes represent key elements driving the partitioning of strain and up-fault fluid flow. These factors are critical in determining oil preservation during the late reactivation phase on the Laminaria High. Testing of the model against leakage indicators defined on 3-D seismic data correlates with the numerical prediction of fault seal effectiveness and explains the complex distribution of paleo- and preserved oil columns in the study area.

Parameterization of elastic stress sensitivity in shales

Stress dependency and anisotropy of dynamic elastic properties of shales is important for a number of geophysical applications, including seismic interpretation, fluid identification, and 4D seismic monitoring. Using SayersKachanov formalism, we developed a new model for transversely isotropic (TI) media that describes stress sensitivity behavior of all five elastic coefficients using four physically meaningful parameters. The model is used to parameterize elastic properties of about 20 shales obtained from laboratory measurements and the literature. The four fitting parameters, namely, specific tangential compliance of a single crack, ratio of normal to tangential compliances, characteristic pressure, and crack orientation anisotropy parameter, show moderate to good correlations with the depth from which the shale was extracted. With increasing depth, the tangential compliance exponentially decreases. The crack orientation anisotropy parameter broadly increases with depth for most of the shales, indicating that cracks are getting more aligned in the bedding plane. The ratio of normal to shear compliance and characteristic pressure decreases with depth to 2500 m and then increases below this to 3600 m. The suggested model allows us to evaluate the stress dependency of all five elastic compliances of a TI medium, even if only some of them are known. This may allow the reconstruction of the stress dependency of all five elastic compliances of a shale from log data, for example.

Enhanced hydrocarbon leakage at fault intersections: an example from the Timor Sea, Northwest Shelf, Australia

Abstract Three-dimensional (3D) numerical modelling of a fault intersection, similar to that suggested being the primary control onhydrocarbon leakage from the Skua Oil Field, Timor Sea, demonstrates that a zone of high dilation can be generated in the vicinity of the intersection during contraction. Only a small amount of deformation was required to initiate these dilational zones, which would probably contain high concentrations of open fractures ideal for high fluid flux in a natural system. The fault intersection was also shown to be an area of relatively low shear strain, which may enhance the potential for fluid flow at these sites due to reduced fault gouge production. Large volumes of hydrocarbons could potentially be lost from these zones of high dilation and low shear where they breach the seal. Therefore, predicting and/or identifying zones of enhanced structural permeability of this type may be critical to accurately assess the integrity of a trap.

FAST: A New Technique for Geomechanical Assessment of the Risk of Reactivation-related Breach of Fault Seals

Scott D. Mildren, Richard R. Hillis, Paul J. Lyon, Jeremy J. Meyer, David N. Dewhurst, Peter J. Boult