Quantifying the Anisotropy of Permeability in Stressed Rock

Lead Research Organisation: University College London
Department Name: Earth Sciences

Abstract

Fluid flow in rocks is vitally important for a wide range of natural processes and human activities, including the triggering of earthquakes, the extraction of oil, gas and water from subsurface reservoirs, and the storage of waste products such as CO2 or radioactive waste. Fluid flow in the Earth's crust takes place through connected networks of pores, cracks and fractures, and is driven by differences in fluid pressure. We measure the ability of rocks to conduct fluid as permeability, and rocks are known to exhibit strong directional variations - or anisotropy - of this key transport property. Laboratory experiments and in situ borehole tests have shown that permeability can vary by several orders of magnitude - i.e. by factors of 100 or 1000 - in different directions. Permeability is also known to be highly dependent on the stress in the solid rock matrix. Again, finely controlled laboratory tests and rather less well constrained in-situ measurements from the subsurface show this to be the case. A key problem though is that the laboratory tests conducted to date have been conducted under simplified stress conditions which do not match the actual anisotropy of in situ stress within the crust. This makes it very difficult to interpret and apply the published laboratory data to more general geological situations, such as fluid flow around seismically active fault zones or reducing risks for CO2 storage in fractured porous reservoirs, with any degree of confidence.

Our proposal is to use a new apparatus at UCL which can apply fully anisotropic (truly triaxial) stress to fluid saturated rock samples of sandstone and granite. Cubic or rectangular shaped blocks of rock will be compressed by three pairs of metal rams, symmetrically arranged at 90 degrees to each other around the sample. This will allow us to vary each of the 3 main (principal) stresses independently. Rock samples will be large enough (5 x 5 x 5 cm cubes, for example) to contain quasi-homogeneous distributions of pores and cracks. We will modify this unique apparatus to enable measurement of permeability along any of the three loading directions that compress the rock. Our proposal builds on recent award-winning research at Aberdeen, where permeability anisotropy has been measured in on oriented samples from a natural fault zone, and carefully related to the pore fabric within the rock. We aim to link the anisotropy of permeability with the anisotropy of stress and the anisotropy of the void space (= pores + cracks). We will define new empirical equations from our quantitative laboratory tests and porosity characterisations. These data and relationships will be used in state-of-the-art computer models of fault zones to explore how directional variations in fluid flow (permeability anisotropy) affect the probability and the type of slip events expected along a fault zone. This will provide a much improved understanding of the risks from earthquake-prone faults in the crust, and more generally, we will begin to understand the truly 3D nature of fluid flow in rocks.

Planned Impact

The following communities of end users could potentially benefit from the proposed research:
1. companies and agencies involved in extracting or storing fluids in the subsurface, including:
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 storing radioactive waste underground, with concerns over the effects of short- and long-term fluid flow in the surrounding rocks

2. agencies responsible for assessing hazards and risks from earthquakes and possibly even volcanic eruptions
a. national survey bodies and safety organisations concerned with accurate forecasting of potential risks
b. policy-makers concerned with evidence-based framing of policy for public debate

These potential beneficiaries might benefit from the proposed research in the following ways:

1. companies and agencies extracting or storing fluids underground need accurate predictive models; directional variations in the rates of fluid flow - a direct consequence of permeability anisotropy - could have huge consequences for short- and long-term predictions of flow rates and fluid volumes that can be extracted and/or stored; this has implications for cost-effectiveness and safety. Better data and better predictive equations will improve economic performance and safety.

2. agencies concerned with risks from natural hazards need realistic and calibrated data for fluid flow; anisotropy of permeability may significantly change these risk assessments, and that is why our key application domain of the data we will collect is fault zone stability

In terms of timescales, we hope to make significant progress with our Project Partner Dr Frederic Cappa with the earthquake hazard modelling within 2 years of the project starting. The more applied/industrial benefits will be delivered through the PhD students to be recruited as part of the Impact plan: it may take 4-5 years from the project start date before tangible benefits are realised in this domain.

Lastly, our proposal will train two PDRAs, and potentially other 2-3 PhD students through the Impact Plan (projects designed in the Impact workshops, and submitted under the NERC CDT in Oil & Gas, Aberdeen is a core member). These professional scientists will be able to apply their quantitative skills to a range of problems in future employment, to the benefit of UK plc.
 
Description In the first few years of this project, we have found that brittle cracking - which eventually leads to failure and faulting - is dependent on all three principal stresses. In any one direction, this has been known for some time as the Kaiser effect. But we have shown (e.g. a published paper in the Journal of Geophysical Research) that it also happens under truly triaxial stresses where all 3 principal stresses are unequal.

A second paper, published in the Journal of Geophysical Research, demonstrates that under truly triaxial stresses that the onset of crack damage is a function of increasing differential stress, regardless of whether mean stress is increasing or decreasing. It showed that the intermediate principal stress then controls the amount of crack damage that ensues and preferentially aligns cracks, but mean crack size remains the same. Cyclic stressing shows that the transition from elastic to inelastic behavior (damage envelope) is dynamic and evolves with stress level

We are currently following up this work by relating it to the anisotropy of permeability in cracked rock which is currently in prep.
Exploitation Route To be assessed at the end of the project.
Sectors Energy,Other