Hydraulic fractures are traditionally modeled as planar features developed by the tensile failure of the rock. Laboratory nanoseismic and field mine-back studies show that most of the fractures are non-planar complex features. Fracture properties are strongly affected by the magnitudes and directions of the stresses in the formation. Low stresses are associated with a complex fracture development while high stresses create simpler, straighter and more planar fractures. We report the results of controlled laboratory triaxial hydraulic fracturing experiments instrumented with piezoelectric sensors. We performed tests on Lyons sandstone which was determined to have an initially isotropic velocity structure. The fracturing experiments have been performed under triaxial stress state to replicate the insitu stress reservoir conditions. The uncertainty in hypocenter locations, frequency analysis, source mechanisms and the effects of stress on fracture propagation will be discussed. Microscopic observations of the fractures are correlated with the mapped microseismic events. Fractures are observed to be non-planar visually and at the SEM scale. Shear failure recorded by focal mechanisms appears to dominate the failure mode. The deviation from planarity will surely affect proppant transport and dispersement.
Surface seismic offers a promising technique to monitor CO2 flood fronts during enhanced oil recovery process. Changes in seismic signature have been observed with CO2 flooding but quantification of the seismic signature with respect to subsurface saturation is still in its infancy. This study is focused on quantification of the variation in seismic parameters (velocity and impedance) with the change in subsurface fluid type and saturation.
The results of a laboratory study are presented where velocity and density were monitored as the pore fluids (formation brine and oil, and CO2) are replaced sequentially. All the experiments were performed at in-situ pressure conditions on plugs (Tuscaloosa sandstones) recovered from a well in a field currently undergoing CO2 flooding. The plugs used are characterized as fluvial (quartz~87%, clay~10%) and distributary channels (quartz~75%, clay~17%).
During brine flooding on dry samples, a decrease in P-wave velocity (~2%) was observed till 95% saturation and thereafter the velocity increases by 15% during the remaining 5% saturation. After attaining 100% brine saturation, oil was pumped to displace brine till irreducible water saturation was achieved. A linear drop of 4% in velocity was observed during this step. Liquid CO2 was injected to displace oil-brine system and a drop of 8% in P-velocity was observed. Associated changes in P-wave impedance due to change in pore fluid saturation are 25%, -5% and -8% respectively for the three flooding experiment. Biot-Gassmann modeling shows good agreement with experimental results for gas-brine and oil-brine system but not for liquid CO2 flooding.
4D seismic data set acquired over the same region is quantitatively interpreted based on these laboratory measurements.