Fourth International Conference on Fault and Top Seals

In this paper, we present the methodology and results of the various fault modelling that was conducted so that the project could confidently proceed from the safety and integrity perspective. The work is broadly divided into two components. The first component is to assess the likely pressures generated by CO2 injection and to assess whether or not the faults would remain stable under such perturbations. The second component of the work assesses both the across-fault and up-fault hydraulic properties of the fault that is proximal to the injection well. Subsequently, the fault parameters are incorporated into dynamic simulations to determine how far vertically up the fault CO2 is likely to migrate. The incorporation of fault properties into commercially available simulator is not a straight forward workflow, and we will discuss approaches that can be taken in this regard.

Temporal fault analysis can give previously unexplored insights into the sealing and leaking capacity of faults over geological time and offer robust estimations of the timing and kinematics of fault histories. A fundamental aspect of temporal fault analysis, particularly quantitative fault seal studies, involves the sequential incorporation of decompaction and valid structural restoration to produce accurate representations of fault and horizon palaeo-geometries. This study uses Midland Valley’s Move™ software to detail the differences seen in fault displacements analyses when compaction is taken into account. Real and synthetic models have been used to highlight the importance of decompaction in displacement and seal analyses. Magnitudes of displacement are shown to be underestimated by up to several hundred metres when decompaction is not carried out and periods of reactivation on a fault can be obscured. Furthermore, estimation of temporal shale gouge ratios and palaeo-juxtapositions are likely to be more realistic with the use of sequential restoration, accounting for sediment compaction. We propose a sequential restoration and decompaction methodology as a fundamental part of any temporal fault displacement and fault seal analysis.

Faults and associated fault zones are key controls on many processes within the brittle upper crust; their individual properties are a controlling factor on the mechanical and fluid transport properties of their protoliths. Despite an impressive volume of work ranging from classic concepts to recent key advances, many features of fault structure and along-strike fault heterogeneity are poorly constrained. This knowledge gap can partially be attributed to outcrop studies that are hindered by a lack of continuous, truly three-dimensional exposures. This work aims to address this paucity in data and presents the preliminary results from an unparalleled dataset comprising LiDAR, photogrammetrically-derived point clouds, wireline geophysics, coal-seam survey data and field observations of fault, and fault-zone architectures. We aim to present preliminary results from a study of fault, and fault zone architecture within mixed siliciclastic and carbonate rocks contained within the Midland Valley basin, Scotland, UK. Three-dimensional fault architecture models have been created using LiDAR, photogrammetrically-derived point clouds and surveyed coal data interpreted in a three-dimensional virtual environment. Models are augmented with measured fault core geotechnical data in order to accurately numerically simulate fluid flow.

We presents results from an experimental program designed to shed light on the effect of stress and deformation history on the permeability and slip behaviour of faults in carbonate rocks. We investigate the mechanical and hydraulic behaviour of experimentally created fault cores and damage zones in natural travertine rock samples and also explore the role of a sealing layer on the frictional and hydraulic response of the rock. Following direct shear testing on the blocks, cylindrical plugs with diameter of 38mm were drilled across the slip surface to be tested in a conventional triaxial configuration monitoring the permeability and frictional behaviour of the samples. The results indicate that the fault cross cutting the sample is acting as seal and its permeability is negatively affected by an increase in mean effective stress; slip on the fault plane does not improve the permeability of the fault. It can be therefore concluded that leakage along an un-cemented carbonate gouge cannot be achieved by movement on the fault plane alone, at least not within the range of slip measureable with our apparatus (; other mechanisms (e.g. cementation of the gouge) need to be explored to assess the possible leaking scenarios in faulted carbonate rocks.

A key factor in the appraisal of a discovery is to establish the full extent of the accumulation using prospect fill scenarios. Typically, only very simplistic fill scenarios are derived in which faults are assumed to be sealing at reservoir-reservoir juxtapositions. In areas of complex faulting, a simple approach can give rise to erroneous volume estimates. Ignoring faults with sealing potential can result in an underestimated prospect evaluation and potentially missed pay. A case study from the Sole Pit Basin, Southern North Sea is presented. The study area is characterised by NW/SE trending faults formed during Mid-Late Jurassic extension that are crossed by NNE trending faults and lineaments. Published methods to convert SGR at reservoir-reservoir juxtaposition to threshold pressure were used to predict maximum column heights supported at the reservoir-reservoir juxtaposition, trap fill and potential leak points out of the discovery. The case study illustrates the significantly improved prospect understanding resulting from a quantitative fault seal analysis. The ability to demonstrate an accurate predictive model helps reduce uncertainty and increase confidence in volumetric calculations, and improves the understanding of hydrocarbon distribution within the discovery. This enables a more targeted appraisal programme, with associated risk, cost and time reductions.

Fault transmissibility derived from lateral connectivity between compartments and changes with depletion was recognised as key subsurface uncertainty in the field. Therefore, structural logging and analyses (fault tip extensions), microstructural and petrophysical property analyses were carried out on cores obtained from 2 wells in the field, in order to assess their impacts on hydrocarbon recovery. Results of the sub-seismic fault characterisation and structural analyses were integrated into a structural modelling and dynamic fault seal project to aid in understanding current production behaviour, and predict possible infill locations. The modelling effort involved using the Petrel 3D Static reservoir models, underpinned by measured fault properties from Rock Deformation Research (RDR, 2000) on Brent and Triassic reservoirs. The result of this work showed that majority of the faults will act as strong baffles in production timescale, however, low, base and high fault permeability and thickness parameters were selected for dynamic simulations. Dynamic simulation sensitivities carried out with the range of potential fault properties bracketed the historical production behaviour observed in the field. Using the historically observed data (production, pressures) the uncertainty ranges for the fault properties were constrained and a suite of history matched models created to allow optimisation of potential infill wells.

The work presented examples of producing post-salt fields in the offshore Republic of Congo to illustrate how fault seal behaviour may change through time. Both examples of "non-sealing" faults on an exploration time-scale behaving as "sealing" faults on a production time-scale, and the converse of sealing faults breaking down during production, are presented. The hypotheses put formard in interpreting the field data are evaluated with simple flow calculations. Firstly, an illustration of the fault sealing behaviour during production shows the feasibility of flow retarding across-fault given likely fault permeabilities. Secondly, an in-house fault permeability algorithm is presented which is used as a test of the hypothesis of fault seal breakdown in a case of up-fault communication between reservoirs. These studies highlight the importance of integrating a wide range of data sources and approaches when assessing reservoir compartmentalisation and communication.