Production assessment of Eagle Ford hydraulically fractured wells

Abstract

This Master's thesis was aimed to assess the use of hybrid DFN and dual-porosity models on the production performance of a well pad that is completed with multi-stage hydraulic fractures in the Eagle Ford. The reason resides in modeling in the industry sometimes the coupled geomechanics and fluid flow models lack the presence of the hydraulic propagation network. In this research model, a simplified discrete fracture network was implemented in dual-porosity model. Moreover, the geomechanical component was observed using the Mohr-Coulomb Failure Criterion. Finally, the model results were compared with a model that was created for the same well pad by [Curnow_Thesis], which considered the hydraulic fractures as symmetric planar bi-wing fractures that are placed in pre-determined locations. Three hydraulic fracture stages of two wells from a well pad in Eagle Ford oil window in McMullen County, Texas were used in the modeling study. Moreover, synthetic fluid composition and initial reservoir pressure from published data were utilized. To establish the hybrid DFN and dual-porosity model , the grids of the model were created to combine the predetermined location planar hydraulic fractures numerical model that was initiated by Curnow's [Curnow_Thesis] and Suppachoknirun's [Suppachoknirum] complex discrete fracture network models. In the grid creation process, the unstructured grids of the Discrete Fracture Network was converted to structured grids. Then, the main fracture features were transferred to a Cartesian grid system and the grids were rotated to match the features' orientation. Iterative coupling was used to couple the geomechanical model and fluid flow model. Two grid systems were established, the reservoir grids and geomechanical grids, which were set as the same size. The input parameters for the geomechanical model were acquired from the experimental data collected using preserved cores from a nearby well. History matching was not included in this research. The gas rates especially were not consistent, which made the history matching challenging. Moreover, the lack of bottom hole pressure data added another layer of complexity to the history matching process. To predict the production of the wells for ten-year period, a production forecast was made. The hyperbolic trend was utilized to fit the data following the industry standards for unconventional reservoirs. The cumulative oil production for the three hydraulic fracture stages of well-1 is around 40 Mbbls and 27 Mbbls for well-2 for ten-year period in the DFN models. On the other hand, the ten-year gas cumulative production for the three hydraulic fracture stages of well-1 is 40 MMcf and 25 MMcf for well-2. However, in the planar hydraulic fracture models the rates were less than the DFN model cases. The rate constraints were not achieved by the planar hydrulaic fracture models. There are some conclusions that can drawn from this modeling research study. The simulated bottom hole showed interesting results. Well-1 drainage area is larger than well-2. Thus, during the ten-year period production well-1's production never transmitted to the boundary of the lower permeability barrier. From the geomechanical model, the maximum and minimum stresses were increased due to the poroelastic effect during production. Moreover, a subsidence occurred and the DFN model with low permeability barrier has the highest subsidence value of 0.109 ft. A comparison with the base model[Curnow_Thesis] was conducted. The different hydraulic fracture patterns from the base model resulted higher cumulative production than this research study. This is due to the several different input data that was used in the models.