Unconventional operators require geomechanical reservoir characterization to improve well performance. Fit-for purpose solutions are needed to balance the accuracy of the solution and the time and cost involved to deliver products to ongoing operations. We show an example of the insight provided by analytical GM models constructed from well and seismic 3D data.
A company operating in the Montney play has experienced casing deformation in horizontal wells after hydraulic fracture stimulation. The occurrence of the casing deformation varies with location and formation and is primarily observed close to faults. A comprehensive dataset is available for the field that includes 3D seismic, multi-finger caliper data of the casing deformation, dipole sonic logs, image logs and mini-fracture data, making it an ideal candidate for investigating both the causation of the casing failure and the potential to use the 3D seismic to predict the risk of casing failure pre-drill.
The cause of the casing failure was postulated to be reactivation of pre-existing faults and fractures during fracture stimulation. Fault reactivation has been inferred to occur in the region based on the induced seismicity occurring during fracture stimulation treatments (Bao & Eaton, 2016). The risk and slip sense of such reactivation can be calculated using a geomechanical model (Mildren et. al., 2005). Consequently, calibrated 1D geomechanical models were constructed in order to test this hypothesis. The model was constructed on a vertical pilot well with two subsequent laterals. Both laterals were fracture stimulated and are drilled in opposite directions, sub-parallel to a fault. Casing deformation was observed in one lateral when milling out the plugs, while no such difficulties were encountered on the second lateral.
The 1D geomechanical model accurately predicted the slip sense observed in the well with casing deformation and also the high risk of reactivation. However, the model did not explain the difference in the occurrence of deformation between the two laterals. Consequently, a 3D analytical geomechanical model was constructed using a workflow establishing a consistent property model between the 1D geomechanical model and the available seismic inversion results in order to investigate any variation in risk along the laterals away from the vertical calibration well.
The lateral with observed casing deformation was drilled in a zone of higher differential stress away from the vertical pilot, while the unaffected lateral was drilled in a zone of lower differential stress. The variations in differential stress appear to be a control on the risk of reactivation, directly correlating with variations in shear stress, and in turn, the magnitude of any failure event. The greater the stress anisotropy, the greater the magnitude of any of fault/fracture reactivation event and greater the risk of casing deformation. By building static 3D geomechanical models and understanding the variation in both risk of reactivation and maximum shear stress, it is possible to lower the risk of casing deformation pre-drill.