The advances in hydraulic fracturing technology and horizontal well completions have led in recent years to rapid rise in exploitation and development of tight gas and shale plays all over the world, and particularly in North America. The popularity of new field technology has in fact raised many new questions. In particular, for forecasting the productivity and EUR of multifractured horizontal wells, it is not clear if conventional reservoir simulation concepts can be adapted for modeling or if extra physics must be included to obtain realistic solutions.
This work presents various methods to model multifractured horizontal wells in tight gas sands using a conventional reservoir simulator coupled with geomechanics. Two actual wells in the same formation but fractured with different techniques (i.e., X-link gelled water fracs and un-gelled water (slick water) fracs) are studied. Detailed investigation of the role of fracture conductivity, effects of initial permeability level, net pay thickness, assumed size of the stimulated reservoir volume (SRV), pressure or stress dependent permeability of the SRV and formation, and virgin reservoir were carried out by history matching the rate and cumulative production. It was established that i) history match is not possible without use of stress or pressure dependent permeability and ii) permeability dependence on pressure inside stimulated reservoir volume must be larger than in the rest of the formation. It was also observed that the standard method for using the same geomechanical data both in uncoupled reservoir and coupled geomechanical model will give incorrect results in terms of production.
A new method based on uniaxial deformation theory is proposed to more accurately approximate the geomechanical effects in conventional reservoir simulators without running a fully coupled simulator. The production results from the uncoupled reservoir modeling using the new method for correcting the permeability data for poroelastic effects were remarkably similar to rigorously coupled geomechanical modeling. This work will be of importance to engineers in analyzing and forecasting production performance of multifractured horizontal completions using numerical models. It will allow engineers to use uncoupled (conventional) reservoir modeling as a practical approximation of more complex coupled geomechanical models.
Same wells were used to model the injection process. History matching of the field bottomhole injection pressure using uncoupled and fully coupled geomechanical models showed the importance of including the geomechanics. Sensitivity study of several history matching parameters such as fracture permeability reduction factor, limiting length of fracture propagation and Biot’s constant was performed and its effects on injection pressure were discussed. Possibility of shear failure in the SRV during injection was also studied, using the Mohr-Coulomb failure criterion. No shear failure was detected when intact rock shear strength parameters were used, but significant shear regions were generated when failure envelope represented fractured or weakened rock.
The main contributions of this work are i) better understanding of the role of the geomechanical effects in both production and injection modeling, ii) the demonstration of the need for coupled geomechanical modeling in injection, and iii) presentation of techniques for approximating these effects in uncoupled reservoir simulation. In addition, valuable insights were gained into the mechanics of fracturing and reservoir behavior during production.