Quantifying the Anisotropy of Poroelasticity in Stressed Rock

Lead Research Organisation: University of Aberdeen
Department Name: Sch of Geosciences

Abstract

Rocks in the upper crust of the Earth are often porous, with the pores and cracks filled with fluids like water, oil or gas. Forces acting on these rocks, arising from the weight of the overlying rocks and from plate tectonics, deform the grains and pores and cracks, changing their shape and volume. This deformation occurs before any fracturing or faulting, and is described by a theory called poroelasticity. This theory states that the orientations of the cracks and pores, where the pore fluid resides, exerts a major control on the response of the rock to stress. Fluid-filled parallel cracks occur in patterns around major earthquake prone faults, and these produce a much stronger response than random orientations of cracks or pores. Therefore, the poroelastic properties of rocks are important for our ability to forecast earthquakes on big faults and induced seismicity from human activities such as fluid injection in boreholes for CO2 sequestration or hydraulic fracturing (or 'fracking').

The poroelastic properties of rocks have been measured in the laboratory but all the data measured to date has been under a very special stress condition that probably does not exist in the Earth. Conventional triaxial stress (CTS) applies a vertical stress on a cylindrical rock sample, and then a constant pressure around the sides. We know that the stresses in the Earth vary in all directions, a condition known as true triaxial stress (TTS). And yet we have no poroelastic data from measurements under this stress state.

A newly commissioned apparatus at UCL has been specifically designed to deform fluid saturated rock samples under true triaxial stresses and thus provide a unique and timely opportunity to address the core scientific issues: there are no published measurements of poroelastic coefficients measured under TTS and we urgently need better data to constrain better models of seismic hazard. Recent work by the investigators has shown that TTS produces significantly different patterns and densities of cracks in comparison to similar loading paths under CTS: TTS produces predominantly aligned parallel cracks, whereas CTS tends to produce radial cracks.

We must systematically collect these data under the most likely in situ stress conditions within the crust - true triaxial stress - and we can use these new data to make tested, more robust, models of seismic hazard. Recent work has shown how important crack fabrics are for the fluid pressurisation, and potential weakening, of earthquake-prone faults. Arrays of fault parallel cracks around seismically active faults could produce a short-term fluid pressure change along the fault equal to the fault normal stress, allowing the fault to slip in an earthquake. This has potentially massive consequences assessing earthquake risk on major faults. Married with the increasing demand for accurate predictions of directional variations in stress and strain in the subsurface (e.g. deviated drilling for geothermal energy or hydraulic fracturing), this adds urgency to our rationale.

We will produce open source software from our research, freely available to other scientists, engineers and the wider public. The first tool, currently being tested, will quantify the three-dimensional (3D) patterns of pores and cracks, including their orientations, sizes and shapes. The statistical distributions of these features will be quantified and used to help predict the poroelastic properties using the published theory. The second tool will use our newly measured poroelastic data to revise published models of earthquake triggering. The inclusion of poroelastic deformation in the current models is mixed with the frictional behaviour, but these are very different physical phenomena. Our new code will combine our previous work on the spatial variations of elastic properties around fault zones with the new laboratory measurements to make more robust forecasts of triggered earthquake hazard.

Planned Impact

The communities of end users that will benefit from the proposed research, include companies & agencies involved in extracting or storing resources in the fluid-saturated subsurface, such as: a) conventional oil & gas companies in the UK and beyond, especially those with enhanced oil recovery programmes (injecting water or CO2); b) shale gas companies looking to use hydraulic fracturing (injecting water); c) agencies storing CO2 in underground repositories; d) agencies tasked with storing radioactive waste underground, with concerns over the effects of short- and long-term fluid flow in the surrounding rocks; e) agencies developing geothermal energy in either hot dry rocks (HDR) or hot sedimentary aquifers (HSA). These organisations need accurate predictive models of poroelastic behaviour. Directional variations in the stress and strain - a direct consequence of poroelastic anisotropy - will have huge consequences for short- and long-term predictions of reservoir and cap rock stability, with clear implications for costs and safety.

In addition, we see potential benefits for agencies responsible for assessing hazards and risks from earthquakes and volcanic eruptions including: a) national survey bodies and safety organisations concerned with forecasting of potential risks; b) policy-makers concerned with evidence-based framing of policy for public debate. These bodies need accurate, realistic and calibrated data for subsurface deformation. Anisotropy of poroelasticity may significantly change these risk assessments, and our main application domain is the stability of seismogenic faults. Our proposal will train two PDRAs. These scientists will develop quantitative professional skills that they can subsequently apply in their future employment, to the benefit of UK plc and wider society.

We will engage with the academic community through publications in scientific journals and presentations at international conferences. Some of our submitted papers and conference presentations will be targeted at cross-disciplinary journals and sessions to raise awareness of our work across the broadest possible spectrum of researchers.

We will engage with industry through a workshop to be held in Aberdeen, the centre of the UK oil & gas industry. Induced seismicity has been known in the North Sea ever since extraction began: the Ekofisk and Valhall fields being prime examples. We will convene a 2-day workshop at the King's College conference centre. We will work with national UK industry bodies, e.g. Oil & Gas Authority, to attract a broad spectrum of industry stakeholders. In detail, we will explore how better models of anisotropic poroelasticity, and better forecasting of the risks using our open source software tools, can be leveraged to improve efficiency, reduce costs and improve safety. Session chairs at the workshop will be tasked with leading discussions and capturing the tangible steps needed to effect change. As stability returns to the oil & gas sector, we will look to develop industry-funded PhD projects that can apply our research findings into reservoir engineering and well bore stability.

Engagement with the wider public will be through dedicated channels at Aberdeen. The Public Engagement with Research Unit (PERU) at Aberdeen produce an award-winning range of public-facing events throughout the year to assist academics in engaging audiences with their research. We will work with them to produce content & events for the annual May Festival (over 10,000 members of the public in 2019) and TechFest (Aberdeen), with a focus on school-age children. We will build analogue demonstration models of poroelastic behaviour and the effects on fault slip, using water, sponges and wood. These models will provide visually appealing examples of how understanding the physics and mathematics of rock behaviour can lead to important insights for seismic hazard assessment.
 
Description We propose the generalisation of the anisotropic poroelasticity theory. In consequence, both instantaneous and time-dependent deformation lead to fluid content variations that are different in each set. We present the equations for such phenomena, where the anisotropic properties of both the solid matrix and pore sets are assumed. Novel poroelastic coefficients that relate solid and fluid phases in our extension are proposed, and their physical meaning is determined. To demonstrate the utility of our equations and emphasize the meaning of new coefficients, we perform numerical simulations of a triple-porosity consolidation. These simulations reveal positive pore pressure transients in the drained behaviour of weakly connected pore sets, and these may result in the mechanical weakening of the material.

We attempt to formalise the relationship between the poroelasticity theory and the effective medium theory of micromechanics. We consider a multi-porous extension of anisotropic poroelasticity, where pore fluid pressure may vary within the bulk medium of interest. As a consequence, any inhomogeneities in the material are not necessarily interconnected; instead, they may form isolated pore sets that are described by different poroelastic parameters and fluid pressures. We attempt to incorporate the effective methods inside Biot-like theory and investigate the poroelastic response of various microstructures. We show the cases where such implementation is valid and the others that appear to be questionable. During micromechanical analysis, we derive a particular case of cylindrical transverse isotropy-commonly assumed in conventional laboratory triaxial tests-where the symmetry is induced by sets of aligned cracks.
Exploitation Route Further papers are in development.
Sectors Energy

Environment

Manufacturing

including Industrial Biotechology

 
Title Stress-induced anisotropic poroelasticity in Westerly granite: Mechanical and Ultrasonic data 
Description Data obtained from triaxial loading experiments conducted on thermally cracked Westerly granite which was initially dry and later saturated with water. Text files (.txt) are raw, unprocessed time series of mechanical data including load, shortening and radial and axial strain gauge measurements. Except noted, units are in Volts. Strain gauge voltages can be converted into strain using gauge factor of 2.11 and amplification factor of 40. External shortening can be converted to sample shortening using a machine stiffness of 480 kN/mm. Raw ultrasonic data are combined in folder named BSF*. Data format is binary, and can be read using ASC ltd. software "Insite". Files "sensors.txt" contain ultrasonic sensor positions and time offset corrections. Mechanical and Ultrasonic data are provided for when the sample was dry. Only mechanical data are provided for saturated conditions. 
Type Of Material Database/Collection of data 
Year Produced 2023 
Provided To Others? Yes  
URL https://zenodo.org/record/7788635