Microseismic Monitoring and Geomechanical Modelling of CO2 - bris
Microseismic Monitoring and Geomechanical Modelling of CO2 - bris
Microseismic Monitoring and Geomechanical Modelling of CO2 - bris
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
CHAPTER 9.<br />
CONCLUSIONS<br />
work would be required as well. This example proves that if CO 2 is injected at pressures that are too<br />
large for a particular formation to contain, events with a detectable magnitude will occur, which can<br />
be imaged using a downhole passive seismic array.<br />
Shear wave splitting measurements made on microseismic data are useful as an indicator <strong>of</strong> fractureinduced<br />
anisotropy. This technique was first developed using teleseismic waves with subvertical arrival<br />
angles. The splitting from such waves is relatively easy to interpret, with fast direction corresponding<br />
to fracture strike, <strong>and</strong> splitting magnitude giving the number density <strong>of</strong> fractures. However, microseismic<br />
data has arrival angles which are <strong>of</strong>ten subhorizontal, making the splitting harder to interpret.<br />
Nevertheless, I have developed a technique to invert for fracture properties using rock physics theory,<br />
<strong>and</strong> shown that it is possible to identify fracture orientations using SWS despite a highly unfavourable<br />
source-receiver geometry.<br />
Splitting analysis on the Weyburn events reveals the presence <strong>of</strong> a principal fracture set striking<br />
to the NW, <strong>and</strong> a weaker, poorly imaged set striking to the NE. Previous work on core samples do<br />
confirm the presence <strong>of</strong> conjugate fractures with these orientations. However, there is a discrepancy in<br />
that core analysis indicates that the set striking to the NE should be the dominant set. To underst<strong>and</strong><br />
this discrepancy, <strong>and</strong> to improve the interpretation <strong>of</strong> what the microseismic event locations mean for<br />
storage security, I have constructed geomechanical models to represent the Weyburn reservoir.<br />
The state <strong>of</strong> the art in geomechanical modelling <strong>of</strong> reservoirs is to couple together fluid-flow <strong>and</strong><br />
finite element geomechanical models. Such models can be used to predict the stress evolution during<br />
injection, <strong>and</strong> hence the likelihood <strong>of</strong> brittle failure <strong>and</strong> microseismic events. I have defined several<br />
stress path parameters, <strong>and</strong> have used these to study the controls that reservoir geometry <strong>and</strong> material<br />
properties have on stress evolution. I find that small reservoirs are more prone to stress arching<br />
effects, so long as the overburden is sufficiently stiff. In contrast, flat, extensive reservoirs do not<br />
tend to transfer stress into the overburden. I have found that the extent to which this can happen is<br />
controlled by the smaller <strong>of</strong> a reservoir’s horizontal dimensions. The smaller reservoirs that transfer<br />
more stress into the overburden are more likely to generate fracturing, both inside <strong>and</strong> above the<br />
reservoir. This finding may be an important criterion when selecting potential sites for their carbon<br />
storage potential.<br />
I have also modelled the effects <strong>of</strong> reservoir geometry <strong>and</strong> material properties on the amount <strong>of</strong><br />
surface uplift. The results demonstrate that the amount <strong>of</strong> uplift can differ by orders <strong>of</strong> magnitude<br />
depending on the reservoir geometry <strong>and</strong> material properties. The modelled uplift for the simple<br />
examples ranges from several centimetres, which would be easily detectable, to sub millimetre-scale,<br />
which would not be detectable even in very favourable conditions. This demonstrates the need to use<br />
accurate geomechanical models both when considering the use <strong>of</strong> InSAR as a monitoring tool, <strong>and</strong><br />
also when inverting measured surface deformation for pore pressure change.<br />
To be confident that geomechanical models are providing accurate predictions, it is necessary to<br />
groundtruth <strong>and</strong> calibrate them with observations. There are a number <strong>of</strong> methods that can be used<br />
to do this, one <strong>of</strong> which will be changes to seismic properties. It is known from empirical observation<br />
that stress changes alter seismic properties, <strong>and</strong> that non-hydrostatic stress changes create anisotropy.<br />
I have developed a model that can account for these effects, while being simple to use <strong>and</strong> easy to<br />
170