Final Activity Report Summary - IMAGES (Induced microseismics applications from global earthquake studies)
IMAGES was an initiative between the petroleum industry and global earthquake seismology. The petroleum industry benefitted by the development of tools and techniques for seismic monitoring of gas and oil reservoirs to control rock fracturing, optimise the reservoir production, prolong life of existing reservoirs, and mitigate hazards associated with the occurrence of induced microearthquakes. Earthquake seismologists benefitted by access to high-quality data of seismicity in reservoirs and a unique opportunity (1) to study fluid driven seismicity, in particular, rock fluid interactions and the role of pore-fluid pressure in seismicity pattern and in triggering of earthquakes by fluid flow, and (2) to inspect rupture processes in a field scale and under controlled conditions.
The research involved four case studies of seismic monitoring of hydraulic fracturing. Three involved hydraulic fracture monitoring in Texas (one was in Cotton Valley; two were named Dominic and Davos). The third dataset came from seismic monitoring of hydraulic fracturing at the Soultz geothermal facility. Some key measurements were not acquired for the Dominic dataset; hence researchers developed new techniques for such data limited problems, which are typical in the petroleum industry, and provided novel insight into hydraulic fracturing processes. The Davos and Cotton Valley datasets were more complete: seismic monitoring occurred in two boreholes.
The waveform modelling group investigated the numerical implementation of the first-order ray tracing which is useful for both inversion and forward modeling. A novel analytical technique for ray tracing in anisotropic dissipative media, which is the simplest representation of an anisotropic dissipative sub-surface, was developed.
The inversion for source parameters group examined the possibility of retrieving the moment tensor from borehole data. It was shown that a single vertical monitoring array does not allow inversion of the complete moment tensor. This limitation is of particular concern as hydraulic fracturing may induce seismic events where such inversion is necessary. Thus it was only possible to carry out a limited moment tensor inversion (amplitude and waveform inversions) in the Dominic dataset which had a single monitoring borehole. Further studies showed that the proportion of non-shear slip deformation resolved by the moment tensor inversion decreased as the number of receiver stations increased. Thus the concept that non-shear deformation occurs may be, in part, due to limited receiver networks.
The stress regime and pore pressure in reservoirs group extended the concept that pore pressure diffusion triggers seismicity to take account of the theory that pre-existing cracks are critically stressed. They also developed the theory that asymmetry in the length fracture wings can be used to characterise the lateral gradient of the minimum horizontal in-situ stress.
The seismological analysis group examined interpretational approaches for hydraulic fracturing, by linking seismological analyses with production/reservoir engineering. Key developments were methods to estimate volumetric fracture growth and fluid loss using microseismic data from several case studies and the shear-wave based back azimuth technique which allows high resolution of the induced seismicity.
The velocity model building group focussed on building one-dimensional velocity models and estimating the correlation function of the media to characterise uncertainty: this is a necessary pre-requisite for seismic inversions. The group studied the attenuation from VSP measurements and found that attenuation at the reservoir is negligible. This allowed sonic logs to be related with VSP measurements and the creation of 1-D mean compressional wave and shear wave velocity models of the locality. The group also showed some uncertainties and artefacts caused by neglecting borehole deviation surveys.
The research involved four case studies of seismic monitoring of hydraulic fracturing. Three involved hydraulic fracture monitoring in Texas (one was in Cotton Valley; two were named Dominic and Davos). The third dataset came from seismic monitoring of hydraulic fracturing at the Soultz geothermal facility. Some key measurements were not acquired for the Dominic dataset; hence researchers developed new techniques for such data limited problems, which are typical in the petroleum industry, and provided novel insight into hydraulic fracturing processes. The Davos and Cotton Valley datasets were more complete: seismic monitoring occurred in two boreholes.
The waveform modelling group investigated the numerical implementation of the first-order ray tracing which is useful for both inversion and forward modeling. A novel analytical technique for ray tracing in anisotropic dissipative media, which is the simplest representation of an anisotropic dissipative sub-surface, was developed.
The inversion for source parameters group examined the possibility of retrieving the moment tensor from borehole data. It was shown that a single vertical monitoring array does not allow inversion of the complete moment tensor. This limitation is of particular concern as hydraulic fracturing may induce seismic events where such inversion is necessary. Thus it was only possible to carry out a limited moment tensor inversion (amplitude and waveform inversions) in the Dominic dataset which had a single monitoring borehole. Further studies showed that the proportion of non-shear slip deformation resolved by the moment tensor inversion decreased as the number of receiver stations increased. Thus the concept that non-shear deformation occurs may be, in part, due to limited receiver networks.
The stress regime and pore pressure in reservoirs group extended the concept that pore pressure diffusion triggers seismicity to take account of the theory that pre-existing cracks are critically stressed. They also developed the theory that asymmetry in the length fracture wings can be used to characterise the lateral gradient of the minimum horizontal in-situ stress.
The seismological analysis group examined interpretational approaches for hydraulic fracturing, by linking seismological analyses with production/reservoir engineering. Key developments were methods to estimate volumetric fracture growth and fluid loss using microseismic data from several case studies and the shear-wave based back azimuth technique which allows high resolution of the induced seismicity.
The velocity model building group focussed on building one-dimensional velocity models and estimating the correlation function of the media to characterise uncertainty: this is a necessary pre-requisite for seismic inversions. The group studied the attenuation from VSP measurements and found that attenuation at the reservoir is negligible. This allowed sonic logs to be related with VSP measurements and the creation of 1-D mean compressional wave and shear wave velocity models of the locality. The group also showed some uncertainties and artefacts caused by neglecting borehole deviation surveys.