A New DFM Dynamic Modeling Workflow Through a Non-Intrusive EDFM Method to Quickly Calibrate Fracture Model With Production Data: Practical Application on a Granite Reservoir Case
The presence of multi-scale natural fractures, low matrix porosity and permeability greatly increases the heterogeneity of granite reservoirs, adding significant complexity for reservoir modeling and simulation. Typical DFN models are commonly challenged for its representativity by reservoir engineers, and the DFN upscale simulation method is blamed for low accuracy representing the DFN model by geologists. In this paper, an innovative solution to bridge the gap between geological modeling and reservoir simulation with EDFM (Embedded Discrete Fracture Modeling) method was introduced for the first time, while preserving the complexity of the DFN and avoiding additional run time or steps into the conventional fluid flow simulation workflow. Unlike previous fracture modeling techniques, EDFM method allows to integrate any complex fracture geometry into a reservoir grid model without compromising fracture resolution and with much higher computational efficiency, which makes the quick high-resolution fracture reservoir modeling feasible. This method is introduced into DFN calibration to streamline the DFN to real fracture geometry reservoir simulation and History Matching with real production data. Therefore, a close-loop fracture modeling, calibration and optimization workflow is developed. This workflow includes four steps: 1) Building a Single well DFN and geological model based on the detailed geological data, such as seismic interpretation, borehole image and core analysis of the granite reservoir. The faults and natural fractures were identified, and their geometrical properties were vested. The fractures are characterized into two groups according to their size grades and their hydraulic impact on the fluid flow in the reservoir, which are those large-scale fractures representing highly conductive corridors and those small-scale diffuse fractures with lower overall conductivity and connectivity. 2) Calibration of the DFN through pressure forward modeling to align with well testing data, the fracture number and size could be optimized. 3) Calibration of the fracture properties. In this process, the single well DFN is integrated into reservoir model through EDFM processor and run numerical simulation to perform history matching, the fracture properties will need to be adjusted to agree with the production data. 4) Definition of the DFN properties to bigger range for block DFN model. This workflow was successfully applied on a fractured granite oil reservoir, 3 single wells’ DFN model were built up and calibrated under this workflow. All 3 wells DFN model were matched with the production history very well. And with less additional grids, the computational performance could be 2 to 3 orders of magnitude higher than traditional methods. And with the increasing of the number and the complexity of fractures, much higher computational efficiency could be achieved.
AAPG Datapages/Search and Discovery Article #90350 © 2019 AAPG Annual Convention and Exhibition, San Antonio, Texas, May 19-22, 2019