IMPACT OF HYDRAULIC RESERVOIR STIMULATION ON VELOCITY MODELS USED FOR MICROSEISMIC EVENT LOCATION DETERMINATION

Microseismic monitoring from a nearby well is performed during hydraulic fracture treatments, to record the energy released by rock failure. We typically locate those microseismic events using a velocity model, P- and S-wave arrival times and hodograms. The initial velocity model is created using dipole or cross-dipole sonic logs, and anisotropy is accounted using the Thompsen's parameters. However, since fluids and solid particles are being injected into the zone of interest, the velocity field changes.

We built two simple models to illustrate the impact of the hydraulic fracturing treatment on the velocity model. The velocities used for the first model are from the sonic logs and the anisotropy parameters were determined prior to the hydraulic treatment. A ray tracing algorithm was used to calculate the P-wave and S-wave traveltimes generated by five impulsive sources located downhole in the treatment well. The traveltimes were recorded by an array of eight three-component geophones located downhole in a monitoring well 850 ft away from the treatment well. The second model includes a pseudo-elliptical region, which mimics an area completely bridged and flooded by the fracturing fluid and proppant. This set-up simulates the conditions of the reservoir immediately after the stimulation job has been performed. The center of the ellipse is in the middle of the perforation interval. Its minor axis is approximately 50 ft and its major axis is approximately 650 ft. These dimensions reflect the hydraulic fracture system outlined by the microseismic mapping as well as the forward fracture design. We calculated the P-wave and S-wave traveltimes for each model and compared the differences between them.

We observed that the P-waves arrivals from the second model are delayed, with respect to P-wave arrivals from the first model. In the same fashion, the time difference between P-wave and S-wave arrivals is greater in the fractured model than in the unaltered one. This could be related to the lower velocity values that characterize the fractured zone, which causes the seismic energy to travel slower. We conclude that, if the acoustic properties variations of the formation due to the hydraulic fracturing job are not taken into account during the processing of the microseismic events, their location will have a great level of uncertainty.