• Fractures;
  • Microseismic


The hydrocarbon industry is moving increasingly towards tight sandstone and shale gas resources – reservoirs that require fractures to be produced economically. Therefore, techniques that can identify sets of aligned fractures are becoming more important. Fracture identification is also important in the areas of coal bed methane production, carbon capture and storage (CCS), geothermal energy, nuclear waste storage and mining. In all these settings, stress and pore pressure changes induced by engineering activity can generate or reactivate faults and fractures. P- and S-waves are emitted by such microseismic events, which can be recorded on downhole geophones. The presence of aligned fracture sets generates seismic anisotropy, which can be identified by measuring the splitting of the S-waves emitted by microseismic events. The raypaths of the S-waves will have an arbitrary orientation, controlled by the event and geophone locations, meaning that the anisotropy system may only be partly illuminated by the available arrivals. Therefore to reliably interpret such splitting measurements it is necessary to construct models that compare splitting observations with modelled values, allowing the best fitting rock physics parameters to be determined. Commonly, splitting measurements are inverted for one fracture set and rock fabrics with a vertical axis of symmetry. In this paper we address the challenge of identifying multiple aligned fracture sets using splitting measured on microseismic events.

We analyse data from the Weyburn CCS-EOR reservoir, which is known to have multiple fracture sets, and from a hydraulic fracture stimulation, where it is believed that only one set is present. We make splitting measurements on microseismic data recorded on downhole geophone arrays. Our inversion technique successfully discriminates between the single and multiple fracture cases and in all cases accurately identifies the strikes of fracture sets previously imaged using independent methods (borehole image logs, core samples, microseismic event locations). We also generate a synthetic example to highlight the pitfalls that can be encountered if it is assumed that only one fracture set is present when splitting data are interpreted, when in fact more than one fracture set is contributing to the anisotropy.