2nd EAGE International Conference on Fault and Top Seals - From Pore to Basin Scale 2009

Trap analysis is inherently a three-dimensional problem involving fault geometry, stratigraphy, and seal integrity. The complex interplay of these elements can make it very difficult to understand and visualize the controls on column height – even for a trap analysis expert. Compounding the issue is the uncertainty associated with reservoir distribution, fault offset and seal thickness. Despite the often equivocal nature of the input, fault seal analysis in the industry has historically been done in a deterministic manner with some attempt at varying the stratigraphic model to give low, medium and high cases. Further, the calculation of fault sealing potential has commonly been done on a per fault basis rather than within the greater context of a trap framework. There are, of course, vendor products and practitioners that address column height prediction in an integrated and perhaps stochastic fashion. However, it could be argued that these practices have created a perception that fault seal analysis is a black box affair, which is often perceived as an ‘art’ rather than a rigorous investigation.

Prediction of hydrocarbon column heights and contacts requires simultaneous evaluation of capillary and mechanical bed seal properties together with trap geometry (crest, spill), estimation and quantification of uncertainties, and propagation of these uncertainties through the analysis. We developed a method for making a probabilistic determination of total seal capacity for a hydrocarbon trap, simultaneously considering both bed seal leak mechanisms (capillary entry pressure and hydraulic fracture pressure), and accounting for trap geometry. Assuming unlimited gas and oil charge and uniform seal properties, simultaneous evaluation of top seal capacity and geometric controls on hydrocarbon fill, leads to recognition of only six possible combinations of hydrocarbon fill control in two-phase (gas and oil) systems. Single phase systems are controlled by only one of the three leak mechanisms. Different combinations of top seal and geometric spill controls can achieve very similar final hydrocarbon distributions in hydrocarbon traps. Without investigating all possible controls it is easy to draw incorrect conclusions about controls on hydrocarbon fill in any single accumulation, promoting incorrect prognostications about undrilled traps. When considering field development plans that include pressure support, misunderstanding actual controls on fluid distributions could lead to unanticipated outcomes, such as leakage through the top seal.

A suite of preserved shales from basins widely separated in space and time underwent multi-stage triaxial tests in the laboratory to evaluate static and dynamic mechanical properties. Coupled with these tests was a workflow for extensive characterisation of physical and petrophysical properties, including porosity, bulk density, specific surface area, electrical properties and microstructural evaluation. Good correlations have been found between shale strength (cohesion and unconfined compressive strength) and some physical properties and there are encouraging relationships between dielectric properties and both static and dynamic mechanical properties. Poor relationships were found to friction coefficient and also between properties previously thought good for predicting shale strength, such as P-wave velocity.

Here we present new data on the relative permeability and capillary pressure of fault rocks as well as the stress dependency of relative permeability. The high capillary pressures and low relative permeability of some well lithified cataclastic faults helps explains why some Rotliegend reservoirs in the Southern North Sea are so severely compartmentalised by faults. The low capillary pressure and high relative permeabilities of poorly lithified phyllosilicate-framework fault rocks may explain why it is often possible to obtain good history matches of production data from faulted Brent-type reservoirs without incorporating the multiphase flow properties of fault rocks into simulation models. The multi-phase flow properties of some samples are highly stress dependent even during elastic deformation. These new data suggest that it may be necessary to take into account the stress-dependence of fault rock flow properties when modelling across-fault fluid flow in compartmentalised petroleum reservoirs.

The Nuclear Waste Management Organization in late 2006 initiated geoscientific investigations at the Bruce Nuclear site near Tiverton, Ontario, Canada, to verify suitability for development of a Deep Geologic Repository for long-term management of Low & Intermediate Level Radioactive Waste generated at Ontario Power Generation owned facilities. The Bruce Nuclear site, located on the eastern flank of the Michigan Basin, is underlain by ca. 840 m of Cambrian to Devonian age sediments that lie unconformably on the crystalline Shield. The DGR, as envisioned, would be excavated at a depth of 680 m within the massive Cobourg Formation, an Ordovician argillaceous limestone that is overlain by ca. 200 m of Ordovician shale. Field and laboratory studies conducted as part of a deep borehole drilling, coring, testing and instrumentation program are yielding evidence of an ancient, diffusion dominant groundwater regime that is resilient to repeated glacial perturbations. This evidence includes measurements of extraordinarily low rock mass hydraulic conductivities(≤10-12 m sec-1), matrix porosities (0.01-0.08) and diffusivities (Dp≤10-11 m2 sec-1), and observation of anomalously depressed environmental heads (ca. 250 m) and vertical environmental tracer distributions within the Ordovician sediments. This presentation describes program results relevant to understanding groundwater regime stability and DGR safety.

Comprehensive laboratory and field investigations on the gas transport capacity of Opalinus Clay were conducted in the Mont Terri Underground Rock Laboratory (URL) in the Swiss Jura mountains and in a deep investigation borehole in the Molasse Basin of Northern Switzerland. Field investigations included hydraulic packer tests and gas injection tests in boreholes. The laboratory investigations on Opalinus Clay drillcores from Mont Terri comprised microstructural analyses, determination of the capillary pressure relationship and gas permeability measurements in isostatic cells. In this context advanced techniques were utilised, allowing for 3-D visualization of the microstructure of solid matter down to the nanometer scale. Fluid filled rock samples were investigated with cryo-techniques (shock freezing), which reduced substantially the appearance of artifacts associated with the sample preparation. Thus the focused ion beam-nanotomography (FIB-nt) with a resolution in the order of 10 nm and Transmission Electron Microscopy (TEM) with an expected resolution < 5nm have been applied for 3-D characterization of the pore structure of Opalinus Clay and in particular for the determination of quantitative measures of the pore connectivity.

Capture and Storage of CO2 could decrease global carbon emissions on short time scales. Leakage rates from geological formations are poorly constrained, and are therefore an unknown factor in CCS feasibility studies. Natural CO2 reservoirs in the Colorado Plateau, USA, are analogues for geological storage. In places CO2 has migrated to the surface along fault zones, forming CO2-charged springs and geysers. Recent drilling for oil and water has also provided rapid pathways for CO2 migration. Uranium-series dating of travertine mounds along two fault zones in Utah has provided insight to the timing and rates of leakage along these faults. A continuous record of leakage has been preserved over the last 315ka, with the oldest dated mound at 413ka. The position of leakage has switched through time, and individual pathways have been utilised more than once for flow. Combining dates with volume measurements from these mounds it is possible to calculate the rates and volumes of flow in individual pathways and the long-term time-averaged CO2 leakage rate for the entire system. The observation that leakage of CO2-rich groundwater from a fault can occur for hundreds of thousands of years has implications for geological storage of CO2.

A radioactive waste disposal in a deep clay formation is mainly based on the low, water and gas, permeability of the host rock. This property returns to a low capacity of gases evacuation and hydrogen produced by anoxic corrosion and radiolysis of organic waste, will affect the repository on time. At first raises the phenomenological question at the different space scales: what processes? , what pathways? , what gas pressure? , what couplings? Secondly raises the question of the safety functions: opening of preferential pathways due to gas fracturation?, transfer induced by water displacement due to gas or transfer in gas form? A first answer was brought by Andra in the Dossier 2005-Argile which concluded on the feasibility of the disposal. Since then, Andra has continued (i) researches on characterization of gas migration processes, and (ii) numerical simulation of the production and the migration of gas. The numerical work is covering scales from the disposal cell up to the whole disposal and from operating period to long term post closure. These new results will in term contribute to update a pluriannual program of data acquisitions on fundamental mechanisms of gas production and migration.

Snorre is a leaky commercial oilfield that supports ~300 m oil column in the Tampen Spur area of the North Sea. The field contains undersaturated oil with no gas cap. Pervasive evidence of oil and gas leakage into the cap-rock is well documented (Leith and Fallick 1997, Bond 2000). Oil migration may alter the wetting state of the cap-rock. Today, the reservoir sands are ~12 MPa overpressure. Log derived pore pressure for the Shetland Group mudstone cap-rock indicates a normal compaction state. This pressure disequilibrium between the reservoir and the cap-rock suggest that reservoir overpressure may have post-dated compaction and geologically recent. Phase behaviour models of the petroleum fluids indicate that a gas cap should exist in the trap if the reservoir was normally pressured. The observed oil column is in equilibrium with the average column height potential of the cap-rock estimated from pore size distribution data. This may imply that the cap-rock is at maximum supportable column height today. In addition, subsequent petroleum charge has leaked into the cap-rock possibly at a rate similar to the reservoir charge rate. These data suggest that the Snorre field capped by oil-wet cap-rock is a case example of a dynamic seal.